U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
PB-280 051
Design and Evaluation of
Laboratory Ecological
System Studies
Oregon .State Univ, Corvallis Dept of Fisheries and Wildlife
Prepared! for
Environmental Research Lab, Oulurh, Minn
Dec 77
,'< JL
-------
TECHNICAL REPORT DATA
(Please read laumctiorts on She reverse before comp' '
NO.
EPA-600/3-77-022
2.
. TITLE AND SUBTITLE
Design and Evaluation of Laboratory Ecological System
Studies
AUTHOR(S)
Charles E. Warren and William J. Liss
'ft, &Z0051 ~
ti. REPORT OTkTE ~ ' """
December 1977 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT N<5"
, PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Fisheries and Wildlife
Oak Creek Laboratory of Biology
Oregon State University
Corvallis, Oregon 97331
10. PROGRAM ELEMENT NO.
1BA608
11. CONTRACT/GRANT NO.
Grant No.
R802286
2. 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
14. SPONSORING AGENCY CODE
EPA-600/03
5. SUPPLEMENTARY NOTES
6. ABSTRACT
Design and evaluation of laboratory ecological system studies are considered in
relation to problems and objectives in environmental toxicology. Ecological systems
are defined to be organismic systems together with their level-specific, co-extensive
environmental systems and to occur at individual, population, and multispecies levels
of biological organization. So that the basis for judgments on the relevance and
adequacy of laboratory ecological system studies for solution of problems in environ-
mental toxicology will be clear, a conceptual framework defining with abstract
generalizations the nature of biological systems is presented and employed. And a
graphical calculus is used to deduce isocline systems and dynamic as well as steady-
state behaviors of multispecies systems, so as to illustrate the importance of
empirical evaluation of the capacities, not simply the performances, of laboratory
ecological systems. Within the context of apparent toxicological problems and this
conceptual framework, the relevance and adequacy of laboratory ecological system
studies on toxicant effects and behaviors are evaluated.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
models
environment simulation
closed ecological systems
aquatic biology
communities
populations
pollution
toxicology
laboratory
microcosms
06 C,F,T
14 B
REPRODUCED BY
NATIONAL TECHNICAL
INFORMATION SERVICE
U. S. DEPARTMENT Of COMMERCE
SPRINGFIELD, VA. 22161
IS. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report!
UNCLASSIFIED
21. NO. OF
20. SECURITY CLASS /Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
irOS GOVHWVENT PRINTING CFFlU 1978— 7-»7-l (
-------
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED
FROM THE BEST COPY FURNISHED US BY
THE SPONSORING AGEN.CY. ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE, IT IS BEING RELEASED
IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE.'
-------
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 ia available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/3-77-022
December 1977
DESIGN AND EVALUATION OF
LABORATORY ECOLOGICAL SYSTEM STUDIES
by
Charles E. Warren and William J. Liss
Department of Fisheries and Wildlife
Oak Creek Laboratory of Biology
Oregon State University
Corvallis, Oregon 97331
Grant No. R802286
Project Officer
John Eaton
Chemical Pollutant Section
Environmental Research Laboratory
Duluth, Minnesota 55804
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
DULUTH, MINNESOTA 55804
-------
DISCLAIMER
This report has been reviewed by the Environmental Research Laboratory,
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 constitute 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
—to measure bioaccumulation of pollutants in aquatic organisms
that are consumed by other animals, including man
This report presents a carefully reasoned analysis of the character
of ecosystems as a basis for designing more meaningful., relevant and useful
field and laboratory studies of ecological processes. The potential would
seem to be great for utilizing theoretical frameworks with the breadth of
those presented here to bridge the gap between traditional, overly-simplified
laboratory investigations and responses of intact natural systems.
Donald I. Mount
Director
Environmental Research Laboratory
Duluth, Minnesota
iii
-------
ABSTRACT
Design and evaluation of laboratory ecological system studies are
considered in relation to problems and objectives in environmental toxico-
logy. Ecological systems are defined to be organismic systems together with
their level-specific, co-extensive environmental systems and to occur at
individual, population, and multispecies levels of biological organization.
So that the basis for judgments on the relevance and adequacy of laboratory
ecological system studies for solution of problems in environmental toxi-
cology will be clear, a conceptual framework defining with abstract gener-
alizations the nature of biological systems is presented and employed.
And a graphical calculus is used to deduce isocline systems and dynamic as
well as steady-state behaviors of multispecies systems, so as to illustrate
the importance of empirical evaluation of the capacities, not simply
the performances, of laboratory ecological systems. Within the context
of apparent toxicological problems and this conceptual framework, the
relevance and adequacy of laboratory ecological system studies on toxicant
effects and behaviors are evaluated.
IV
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables x
Conclusions and Recommendations 1
I Problems, Objectives, Methods and Explanation in
Environmental Toxicology 5
II Conceptual Considerations in Environmental Toxicology .... 13
III Theoretical Considerations in Environmental Toxicology ... 29
IV Empirical Considerations in Environmental Toxicology .... 56
V Conceptual, Theoretical, and Empirical Efforts to
Resolve Problems in Environmental Toxicology 92
References 112
-------
FIGURES
Number
1 The conceptual continuum of levels of explanation and
understanding of an individual scientist. This is
considered to be within the context of the scientific
community. After Warren, Allen, and Haefner (MS)
A biological system viewed throughout its existence as
being composed of an organismic system together with its
spatially and temporally co-extensive environmental system.
After Warren, Allen, and Haefner (MS) ............ „ 15
A diagrammatic view of an organismic system showing that from
some potential capacity the organismic system passes through
a particular sequence of realized capacities jointly deter-
mined by its potential capacity and the actually prevailing
environmental system. After Warren, Alien, and Haefner (MS) . 1?
Organismic subsystems and their level-specific environments,
all of which are incorporated by an organismic system having
its own particular environment. After Warren, Allen, and
Haefner (MS) ....................... . . 18
Kinetic diagrams representing defined multispecies system, inclu-
ding plants (P) , herbivores (H) , and carnivores (C) . Rate of
light input (I) and rate of exploitation (E) represent
external environmental variables. Light intensity (R) is a
variable dependent on I and Pc Competition, commensalism,
and mutualism are represented in B. And a toxicant (T) direc-
tly affecting only carnivore C is represented in C ....... 34
Phase planes and interrelated isocline systems representing a
sequence of predator-prey interactions in a simple community. „ 37
Derivation of prey isoclines on the C-H phase plane by graphical
summation of herbivore gain and loss response functions. ... 39
Derivation of the herbivore production gain response functions
used in constructing the prey isoclines on the C-H phase
plane ................ . ............ 41
Derivation of predator isoclines on the C-H phase plane by
graphical summation of carnivore gain and loss curves. . . . . 45
VI
-------
Number Page
10 Derivation of the carnivore production gain response functions
used in constructing the predator isoclines on the C-H
phase plane 47
11 (A) Direct effects of toxicant T on the relative growth
rate of carnivore C in relation to herbivore H biomasses,
at concentrations of OT, IT, and 2T. (B) Direct effects
of toxicant on the loss or mortality rate of carnivore
C are shown to occur at 2T but not at OT and IT and are
taken to be linearly proportional to biomass of C 49.
12 Possible steady-states (indicated by circles, squares, and
triangles) and trajectories of biomasses of carnivore C,
herbivore H, and plant P population and light intensity R
under different rates of light input (low I, medium I, and
high I) and different levels of fishing effort E, when
toxicant T is absent and when it is present at concentra-
tions of IT and 2T 50
13 Steady-state biomass isopleths of the alga Chlamydomonas
reirihardtii (A and B) and the protozoan Tetrahymena vorax
(C and D), under different light and flow regimes when
nitrate concentration was 0.5 mM (A and C) and when it was
0.05 mM (B and D) in continuous flow cultures. These
diagrams illustrate population scopes for performance in
terms of steady-state biomasses and environmental
variables. After Taub (1973) S3
14 Effects of the addition of 0.2, 2, and 20 ppm Aroclor 1242 on
the growth of Chlamydomonas reirihardtii. Each point
represents the mean of two observations whose individual
values are indicated as the range. After Morgan (1972). ... 68
15 Diagrammatic representation of theoretical and empirical
partial evaluation of capacities of laboratory ecological
systems by determination of scopes for performance as sets
of performances under different environmental conditions. ... 73
16 The influences of a controlling factor such as temperature
(A), lethal levels of the controlling factor (B), a limiting
factor such as dissolved oxygen (C), and the cost of
additional physiological regulation as might be caused by a
toxicant (D) on the active and standard metabolic rates and on
the metabolic scope for performance of a hypothetical
poikilothermic animal such as a fish. After Fry (1947) 75
17 Theoretical effects of temperature change on the food consumption,
energy budget, and scope for growth performance of a
hypothetical poikilothermic animal having food available in
different amounts. After Warren (1971) 76
vii
-------
Number Page
18 Relationships between median survival time of rainbow trout
and concentration of zinc ion in water having different
total hardness. After Lloyd (1960) 77
19 Percentage isopleths representing scope for survival of the
fourth zoeal stage of larvae of the crab Sesamu .eiruevewn
in relation to temperature and salinity of water. After
Maguire (1973) 79
20 A three-dimensional diagram showing the influence of both
oxygen concentration and water velocity on the mean dry
weights of newly hatched coho salmon fry developing from
embryos reared throughout development at the various
combinations of oxygen concentration and water velocity
represented by the intersection of the curves. After
Shumway, Warren, and Doudoroff (1964) 80
21 Relationships between rates of food consumption and growth of groups
of steelhead trout kept at different fluctuating tempera-
tures in a spring experiment. Rates of food consumption
and growth are expressed in dry weights. The initial mean
wet weight of the fish was 2.29 g. Numbers in parentheses
indicate the number of fish which died in a treatment.
After Wurtsbaugh (1973) 81
22 Energy budgets showing relationships between food consumption
rate, energy and material uses and losses, and dissolved
oxygen concentration for juvenile coho salmon in labora-
tory studies at 15C in the.summer. The influence of both
food availability and oxygen concentration on scope for
growth are apparent in differences in growth parts of these
budgets. After Thatcher (1975) 82
23 The finite rate of increase (A) as a function of temperature
and moisture for the grain beetles, Calandpa oryzae and
Riizopertha dominiaa living in wheat. Rates of increase
are indicated by contour lines that describe conditions
with identical values of X. After Andrewartha and Birch
(1954) 83
24 Curves indicating relation of yield of guppies per 3-week
brood interval to biomass and exploitation rate (indicated
percentages) at each diet level. Points indicated for
0 percent exploitation rate are average population levels
for the 3 weeks immediately before exploitation. After
Silliman (1968) 84
Vlll
-------
Number Page
25 Approximations of biomass-production curves for cutthroat
trout in the laboratory stream ecosystems having five
different basic capacities to support trout production
in the spring 1968 experiment. After Warren (1971) 86
26 Relationships between food consumption and growth of juvenile
chinook salmon at different concentrations of kraft process
pulp and paper mill effluent, when the fish were fed con-
trolled rations in aquaria (A). And (B) relationships
between the biomass of juvenile chinook salmon and their
production in laboratory streams receiving similar effluent
and in streams not receiving such effluent. After Tokar
(1968) and Seim et al. (1977) 87
27 R elationships between the growth rate of sculpins and the
density or biomass of the insects upon which they fed
in laboratory streams into which different concentrations
o¥ dieldrin were introduced (A). And (B) relationships
between the biomass of sculpins and their production in
the same streams. After Warren (1971) 88
28 Delineation of human goals, problems, institutions, and
institutional objectives and approaches 97
29 Delineation of the goals, problems, objectives, and approaches
of man's scientific institution 98
30 Problems, objectives, and methods of laboratory ecosystem
studies in relation to those of other studies 100
31 Apparent and possible objectives of laboratory ecosystem
studies in environmental toxicology 101
IX
-------
TABLES
Number Page
1 Ecological magnification indices (concentration of chemican in 71
organism/concentration of chemical in water) obtained from
laboratory multispecies systems having different designs and
experimental protocols.
-------
CONCLUSIONS AND RECOMMENDATIONS
1. In the creation and utilization of environments to satisfy human needs
and aspirations, resource, cultural, political and technological problems
are perceived. The introduction of massive amounts of toxic substances for
pest control and as a result of industrial and other processes has led to the
development of such problems. Solutions of these and other problems having
social dimensions are sought through the legislative, judicial, administra-
tive, educational, technological, and scientific institutions societies develop.
This report deals primarily with scientific approaches—particularly laboratory
ecological system studies—to problems associated with effects, transport,
accumulation, and degradation of toxic substances. But the context of use
and success of scientific approaches is the entire complex of social, insti-
tutions, so that not only the limitations of science but also the limitations
of other institutions are involved. ^
2. The problems with which this report is primarily concerned are: accu-
mulation of toxic substances in biological products utilized by man; reduc-
tions in production and yield of biological resources caused by toxic
substances; and toxic alteration of the capacity, structure, and persistence
of ecological systems of importance to man. And the objectives of lab-
oratory ecological system studies with which we deal are taken to be:
assay of the relative toxicity of toxic substances; determination of the
behavior of toxic substances; and determination of the effects of toxic
substances on individual organisms, biological populations, and biological
communities.
3. Science generally recognizes that its methods must be evaluated against
the objectives they are intended to accomplish, but it is not so generally
recognized that the objectives of science ought to be evaluated against prob-
lems of understanding or more concrete problems of life. The difficulty of
evaluating objectives against problems, and even methods against objectives,
is that different people perceive problems and even understand stated objec-
tives differently. This is a natural outcome of differences in conceptual
structures within which people understand their experiences. Such differ-
ences lead to difficulties in all human institutions, including science.
Disagreement ensues as to what the problems really are, what objectives ought
to be, and how these are to be approached. Disagreement in itself is not unde-
sirable but it cannot be dealt with rationally unless its bases are
explicated.
-------
4. Rationality can be taken to be that which makes understanding possible.
Since Immanuel Kant, philosophers of science have recognized that any
understanding whatsoever depends on a priori conceptual constructs. But
these constructs are not usually specified, even in science, at any higher
level than formal theory. Were they to be, understanding, and thus ration-
ality, would more generally be possible, even when there are disagreements over
assumptions and conclusions. We take what we will call a conceptual framework
to be a description of some domain of human experience, together with rules
for its use. Specification of such conceptual frameworks is needed in all
human institutions including science, for they would serve as an important
aid in communication, problem perception, and evaluation of the adequacy of
approaches to problems.
5, In this report we review a conceptual framework developed for biology-
It consists of a set of four abstract generalizations articulated in natural
language, together with conceptual and procedural rules. The generalizations
define a biological system, at any level of organization, as an organismic
system together with its level-specific co-extensive environmental system,
through all states of the organismic system from its origin to its end. We
believe this to be necessary for adequate explanation of biological systems.
Organismic systems exhibit performances like organization, structure, replica-
tion, and persistence. To exhibit such performances, organismic systems
must have the capacities to do so. Performances are determined by capacities
interacting with environments. Thus, so long as environments are changing,
as in nature they must, performances are likely to change, and no particular
time-variant or steady-state performance is of much interest. It is the
capacity of the organismic system that is of interest, and theoretical and
empirical approaches ought to be designed to make partial evaluation of
capacities possible—no full evaluation of a theoretical concept such as
capacity being possible.
6. Conceptual frameworks, in the very general way in which they describe
whole domains, do not provide the particular explanations of particular
systems that we need. For this, theories as abstract scientific deduc-
tive systems leading to theorems that can be translated into empirical
generalizations are needed. Particular empirical generalizations alone do
not account adequately for human observational experience. A universal
theory allows us to deduce many different empirical generalizations and so
explain and logically unify them. In so doing, theory makes more clear the
boundary conditions of empirical generalizations,-which is very
important in any applied science such as environmental toxicology. But
because of its greater particularity, a theory leaves out. much of what must
be involved in natural and social systems, this making conceptual frameworks
as well as theories necessary in human understanding and solution of problems.
7.. For defined organismic systems, we illustrate that changes in the number
and levels of environmental factors, including toxicants, change not only
the time-variant behavior of the systems and their parts but also their
steady-states. No particular observation of a behavior that is time-variant
or even in steady state is of much interest. Rather, it is the domain of
possible time-variant behaviors and steady states under ranges of environmental
-------
factors that is of more general interest and greater value in application
to solution of problems in environmental toxicology. With changes in
process and structure of organismic systems on any level of biological
organization, changes in the effects, transport, accumulation, and degradation
of toxic substances are to be expected. Thus any particular measurement of
these may not be of much value in application to natural systems and the
solution of problems occurring there.
8. Organismic systems can be defined as occurring at individual organism,
population, and community levels of biological organization. Laboratory
models of ecological systems can be created at the individual level,
the population level, and sometimes at multispecies levels, but probably
not at the level of the biological community. Individual organisms, because
of evolution, reproduction, and development, are natural organismic systems
having their species-specific natural capacities. Populations developing in
the laboratory through the reproduction, development, growth, and survival
of individual organisms may have capacities resembling those of natural
populations of the same species, but these capacities are likely to be
influenced by laboratory environmental conditions. Multispecies systems
created in the laboratory generally do not have capacities of natural multi-
species systems. This is because in bringing together different species
having their own capacities, we create in the laboratory higher level
systems having capacities in part determined by our own actions. It is not
clear to us that multispecies systems having capacities for organization,
structure, and persistence have usually been created in the laboratory.
But even when they may have been, it is difficult to know what if any
natural level of biological organization has been modeled, unless some
carefully studied natural multispecies system has been recreated in the
laboratory. This has not often been done. Whatever is created in the
laboratory, it must generally be at a much lower level of organization than
a biological community, which together with its environment forms what is
usually called an ecosystem. And we ought not to attribute directly
to natural ecosystems performance and capacity evaluations based on model
ecosystems. This applies to any sort of evaluation of the effects or behaviors
of toxic'substances.
9. Individual organisms, because of their naturally endowed capacities,
can generally be more meaningfully evaluated in the laboratory than can
higher levels of organization. Moreover, it is generally more practical to
evaluate these capacities through determination of scope for performance
over a range of kinds and levels of environmental factors. This applies to
effects and behaviors of toxicants, as well as to other environmental factors.
Such evaluations are most valuable if the entire life history of the individual
organism is taken into account.
10. Development of a laboratory population, as a higher level system,
requires provision of environmental conditions suitable for completion of
individual life histories and persistence of the developing population
through several generations. Creation and study of population systems in
the laboratory is practical only for small organisms having relatively
short life histories. -When it is accomplished, the performances and
-------
capacities of laboratory populations should be evaluated over as wide a
range of kinds and levels of environmental conditions as possible, especially
if results with toxicants and other environmental factors are intended for
application to solution of problems occurring in nature.
11. Multispecies systems having their own level-specific capacities are
very difficult to develop in the laboratory. This requires not only the
provision of conditions suitable for life history completion and individual
population persistence but also that the populations are so adapted one to
another as to form a multispecies system that as a whole has the capacity
to persist under provided environmental conditions„ This has rarely been
done for systems of more than two species of small organisms. And even if
such systems are successfully created, it is not any particular time-variant
or steady-state performance that has much meaning. Their domains of time-
variant behaviors and steady states should be evaluated, even if only partially,
by study of the system under as wide a range of kinds and levels of environ-
mental factors as possible. This is important if we are interested in
effects, transport, accumulation, and degradation of toxic substances, because
these will all change as functions of system process and structure. Some
problems may best be approached by means of multispecies laboratory
systems. But, if this is the intention, there must be a commitment
to the exceedingly difficult and long-term process of creating and
studying such systems. Anything less is not likely to be worthwhile
and may be very misleading.
12. Not only the perception of problems but also the different inter-
pretations given even defined objectives demand some specification of
the conceptual structures on which these are based, if our search for
solutions to problems in environmental toxicology is to be made under-
standable and thus rational. Moreover, the design and interpretation
of the sorts of experimental studies of ecological systems considered
here are dependent on presuppositions, whether or not these are speci-
fied. But without development and specification of conceptual frameworks
and theories, the design of experiments and the meaning of their results
can hardly be evaluated, at least in a rational way that can be commu-
nicated to all concerned. And finally, ways of applying results to nature
for the solutions of problems in environmental toxicology are never
implicit in the experimental results themselves, because the systems on
which these were determined are much too different from nature. Here
again, and for all these reasons, conceptual frameworks and theories
are needed. Far more effort in environmental toxicology should be
devoted to the development of adequate conceptual frameworks and
theories than has been in the past. Understanding and solution of
our problems with toxic substances can never be based on empirical
investigation alone.
-------
I. PROBLEMS, OBJECTIVES, METHODS, AND EXPLANATION IN ENVIRONMENTAL TOXICOLOGY
A. PROBLEMS AND OBJECTIVES IN ENVIRONMENTAL TOXICOLOGY
Any .consideration of the design and evaluation of laboratory ecosystem
studies in environmental toxicology should deal with social-technological
problems as well as with objectives and methods of research intended to be
useful in the reduction of these problems. For the apparent relevance of
objectives and the adequacy of methods in environmental toxicology depends
very much of the nature of the problems, or, more precisely, on how we
perceive these problems. Those of us interested in environmental toxicology
know something about problems associated with the use of toxic materials.
And we are aware that the solution of such problems demands socio-economic
and technological knowledge and action that would best be based on scientific
understanding. But we are also aware that social solutions may need to be
attempted, and very often are, in the absence of adequate understanding.
In consequence, continued and uncontrolled use of some very hazardous
materials may be permitted. And severe restrictions may unnecessarily be
placed on the use of other materials having considerable potential value
to man.
But in all this, there is very often sincere disagreement among
individuals and groups as to just what the problems really are and whether
or not the uses of particular toxic materials present unacceptable hazards.
Differences in the interests and knowledge of people inevitably lead to
differences in problem perception. The occurrence of differences in problem
perception and the social and scientific difficulties to which this leads
should, at the outset, be clearly recognized. We are dealing with social
and natural problems that are probably of irreducible complexity, and there
is certainly much to justify different opinions. Differences of opinion,
in themselves, are not bad and indeed may lead to elucidation of different
aspects of complex social and natural systems. But rational consideration
of differences of opinion is only possible if their conceptual and obser-
vational bases are made known, and this is rarely done. The need for
social action, the relevance of scientific objectives, and the adequacy of
scientific methods then become matters in environmental toxicology that
are very difficult to evaluate.
The problems with which we will here be concerned result from the
introduction of relatively large amounts of toxic materials into the
environment; we will not consider problems of drug toxicity or occupational
health. Problems associated with pesticide use in agriculture and forestry
prompted preparation of this report, but most of the ensuing discussion
pertains as much to toxicants in industrial effluents. Along with benefits
deriving from the use of pesticides, there are many side effects, not all
of which are known but many of which have undesirable environmental effects
and may be quite hazardous to man himself. Reduction or loss of natural
populations, alteration in the structure of biological communities,
decreased production and^yields of exploited wild populations, and the
accumulation of substances hazardous to man are among the problems that
concern us. The biological systems as well as the social systems within
which such problems are perceived to occur are almost unknowably complex.
-------
They extend from physiological levels of organization to biological commu-
nity levels and even beyond. Here, we will mainly address ourselves to
scientific approaches dealing with individual organism, population, and
higher levels of biological organization.
We take the ultimate goal of science to be the continuous advance of
human understanding. Science does this through conceptual frameworks
and theoretical and empirical explanations. And explanation, so as to
achieve more understanding of phenomena in environmental toxicology,
ought to be the primary objective of investigation of toxic materials.
But understanding of complex biological systems will ever be "very partial,
and scientists, recognizing this, generally have very limited objectives:
assay of the relative toxicity of different materials, determining their
behavior in biological systems, and determining some of their effects on
individual organisms, populations, and perhaps even biological communities.
So as not to appear naive, we often claim for our approaches to such
objectives only that they are bioassay techniques, or other ways of estab-
lishing environmental standards, or screening techniques, or simple models
to give us at least some initial understanding. But in so doing, we
ought not to delude ourselves that our findings thus need not relate to
problems in complex natural and social systems. For it is to these
systems that the results will be applied. A simple laboratory technique
for screening for approval of use of toxic materials that does not account
for natural complexity may lead to approval of a hazardous material, or
needlessly prevent the use of some potentially valuable one.
Thus our objectives as well as our methods need our continuous evalu-
ation. Scientists generally recognize that evaluation of their methods
is always legitimate but they may resist evaluation of their objectives,
this seeming to infringe upon scientific freedom. We can accept the need
for a diversity of objectives in any vigorous science and still recognize
that all objectives are not likely to be equally relevant or worthwhile
to the pursuit of solution of particular problems. Objectives, in our
opinion, should be scrutinized in view of the nature of the problems to
which they are claimed to pertain. This is no simple matter, because so
doing requires that we have some at least initially adequate notion of
the nature of the biological system in which a problem appears to occur.
A major section of this report will be devoted to consideration of the
nature of biological systems, in part to facilitate evaluation of
objectives,,
Science might be described as the art of defining problems in ways
that they can be solved. The greatest advances in physics have been
based on this. Still, this way of defining, or redefining, problems in
environmental toxicology could easily lead to superficial solutions the
application of which could be hazardous. Biological systems have never
been shown to be amenable to easy definition or representation.
Finally, one can hardly evaluate objectives that are not clearly
stated. It is not easy for us to define well our objectives, especially
-------
for very exploratory studies. But the guidance clear objectives give us
continues to make the effort important. And clearly stated objectives
are absolutely essential if others are to be able to evaluate our methods.
B. EVALUATION OF METHODS IN ENVIRONMENTAL TOXICOLOGY
Even when objectives of biological research are clearly set forth,
the evaluation of methods intended to make possible the achievement
of these objectives is in no way a simple matter. In part, this is
so because individual scientists differ in their views of the nature of
biological systems and of what it would require to explain and so advance
understanding of these systems. In consequence, except for the most
limited objectives, biologists are apt to disagree as to whether or not
particular methodological approaches are adequate. And because biological
systems are complex, their adequate explanation is generally possible
only through articulation of their different aspects, each of which may
reasonably be approached methodologically in different ways. There is,
then, considerable justification for a diversity of methodological
approaches even to the same general objective.
In later sections, we will set forth our views of the nature of bio-
logical systems at conceptual framework and theoretical levels of under-
standing. For now, let us comment only briefly on three aspects of the
evaluation of laboratory methodology: representational adequacy of the
designed system, the adequacy of measurements, and the problem of inter-
pretation. If we are to employ laboratory ecosystems, then represen-
tational adequacy, at the empirical level, pertains to whether or not the
designed systems represent important aspects of natural ecosystems suf-
ficiently well to make possible achievement of our objectives related to
.the natural systems. Studies of an intact living organism in the labora-
tory, however much we may modify its environment, are at least studies of
a system that has acquired its design through evolution and development:
the organism, itself, is a natural system and is represen -
tationally adequate for studies directed toward some important objectives.
But it is another matter altogether where a biologist brings together in
the laboratory groups of individuals of different species, maintains•them
under certain physical conditions, and begins to think of such an artifact
as a laboratory ecosystem. Unless such a system is put together in light
of considerable knowledge of some natural system it is intended to mimic,
we have no reason to suppose that its measured behavior will be of any
interest at all. Even when such knowledge is employed, the designed
system is still likely to be dimensionally, relationally, and dynamically
inadequate for attainment of objectives related to the natural system of
interest.
Any laboratory ecosystem designed to be representationally adequate
will have dynamic behavior that is determined by the number and kinds of
species populations, their interrelationships, and environmental conditions.
In its dynamic behavior, the values of its biological and physical variables
may tend toward system steady states so long as environmental conditions
are fixed. But even ,such a well designed laboratory ecosystem—of which
-------
we are aware of very few—has many potential steady states, which are
determined by the levels of environmental variables. With changing
environmental conditions, the system will not long persist in a given
steady state,, Measurements of system variables, then, can within broad
limits be of almost any values, according to the state of the system when
the measurements are made. These measurements represent time-variant
behavior of the designed system, not simply variance to be dealt with
statistically. If the system is not so well designed, measurements of
its time-variant behavior may be of no interest whatsoever. Even good
replicates of a well designed system are likely to be out of phase in
their dynamic behavior, and the meaning of results determined only at one
or a few times becomes difficult to ascertain.
This raises the whole question of how measurements on even very -well
designed laboratory ecosystems are to be interpreted. For we ought to
know that had the measurements been made at a different time they in all
probability would have been different. Moreover, were the levels of
environmental factors to have been different, another pattern of dynamic
behavior would then have occurred. And, of course, were one or more
additional dimensions to have been included in the system there would
probably have been completely different patterns of behavior. The interpre-
tation of measurements requires that we have some a priori view that can
give them meaning. Such a view may be held only subconsciously, but it must
be there, as important philosophers of science have agreed.
We make measurements and we give them meaning: nature does neither
of these for us0 Moreover, when we design laboratory ecosystems, we in
large part determine the capacities of these systems, their possible
behaviors, and thus what measurements are possible. All this raises
exceedingly difficult questions as to how we are to give meaning to the
results of laboratory research on biological systems above the individual
level of organization. Empirical generalizations, based on observational
experience, give meaning only to the experience upon which they were
based. We can know little about their generality—little about the
boundary conditions of their application to nature.
Theory is another methodological approach science employs to give
meaning to observational experience. From adequate theory, we should be
able to. deduce, and so explain,.sets of important empirical generalizations.
This tells us much more about the generality and applicability of empirical
results. Moreover, however adequate a theory, much of its meaning derives
from higher level conceptual constructs: the systems of consciously or
subconsciously held beliefs about the nature of natural systems. For it
is not only observational experience that determines what theories we
find acceptable. These higher level conceptual constructs we will term
conceptual frameworks, and we will distinguish clearly between these and
the formal scientific deductive systems we call theories. In a later
section, we will present a conceptual framework for biological
.systems, because we believe this and other such frameworks have important
roles to play in empirical as well as in theoretical investigations. We
believe that they can do this best if they are made explicit.
8
-------
C. EXPLANATION AND UNDERSTANDING IN ENVIRONMENTAL TOXICOLOGY
Environmental toxicology, as a domain of human concern and applied
science, is not unified by any common view or well articulated body of
scientific knowledge. The interest groups and scientific disciplines
involved are very diverse. Its being an area of common concern at all
derives mainly from acknowledgment that toxic materials are being intro-
duced into our environment, that sometimes the reasons for this appear
very good, but that there are known and unknown hazards associated. And
the full complexity of the matter is hardly perceived,, Various biological
and physical sciences are involved, but there is no general theory of
toxicology—even at physiological levels, to say nothing about ecological
levels of organization. There is some pertinent physical theory, however
adequate it may be, but in toxicology we must finally deal with biological
systems„ The apparent goal of environmental toxicology would seem to be
evaluation of the behavior and effects of toxic materials in biological
systems, so as to make possible appropriate use and control of such
materials. But just what such evaluation would ultimately need to entail
is not at all clear. Mainly the effort until now has been empirical,
and has proceeded through an increasing array of techniques providing
results that are sometimes difficult to articulate and apply with any
certainty at all. We have little general explanation and understanding of
the phenomena with which we must deal. And without this our successes in
dealing with toxic materials are likely to continue to be too few and too
partial.
We suppose the goal of science to be the continuous advance of human
understanding. Science works toward this goal by creating different
sorts of explanations that make comprehensible and logically unify as
much human perceptual experience as possible. Any explanation, most
generally, can be taken to be a translation of the unfamiliar to the
familiar that satisfies the need for explanation. To be an explanation,
a statement or set of statements must increase understanding. But
familiarity and satisfaction are relative. A language, or a statement
articulated in a language, that is familiar to one person may be foreign
to another„ And even a familiar statement that, as an explanation,
satisfies some may not satisfy others. Understanding is not advanced
by proffered explanations alone but rather by these in interaction with
the conceptual context within which they are received. Some commonality
of language and conceptual constitution are a pri-ori requirements for
explanation and the advance of understanding.
Now to some the foregoing may seem rather far from environmental
toxicology and the design and evaluation of laboratory ecological system
studies. But we think it is important in this endeavor, because it lies
at the roots of the possibility of scientific understanding and thus must
also be the basis of environmental toxicology. We are here primarily
concerned with the meaning of laboratory results and the application of
these in the resolution of environmental problems„ And man,, not nature,
gives meaning to his observational experience. How he does this is what
science is all about.
-------
To the early logical positivists, most notably Rudolf Carnap, the
meaning of an explanatory proposition was its "method of verification."
This was generally taken to be a deductive argument leading to logical
consequences that could be related to observational experience, which was
thus explained. Internal consistency and predictive power became the
dominant criteria of explanatory adequacy, little attention being directed
to conceptual presuppositions and other aspects of "external" adequacy of
explanation. Nagel (1961), Popper (1959), and other philosophers
recognized that theoretical constructs used in explanations do not in. any
simple way emerge from observational experience—that they are instead
products of the scientific imagination. But these philosophers persisted
in emphasizing internal consistency and verification or falsification by
means of observational experience as the principal if not only criteria
of explanatory adequacy. Some philosophers of science, certainly Kant,
but more recently Hanson (1958), have insisted that the meaning
of explanations comes as much from their encompassing contexts of precon-
ceptions, beliefs, and background of knowledge of the persons involved as
from the internal consistency and observational correspondence criteria
emphasized by the logical positivists« Indeed, the meanings we associate
with particular observational experience, say laboratory results, and
thus the uses of results we are comfortable in making are equally dependent
on such encompassing conceptual contexts.
In dealing with these philosophical matters pertaining to the conduct
of scientific investigations, Warren, Allen, and Haefner (MS) employed a
notion of the "conceptual continuum" of the individual scientist, which
is illustrated diagramatically in Figure 10 This representation suggests
that there are different levels of explanation and understanding extending
from observation statements through empirical generalizations, models,
theories, and conceptual frameworks to our funds of accruing experience
and beliefs about the living world. Implicit in this is our conviction
that observational experience and explanation at any level of such a
continuum are understood only within the context of the whole. Thus the
meaning we associate with results obtained from laboratory models—or any
meaning we would be able to associate with such results—-depends very
much on available and operative theories and conceptual frameworks, whether
or not these are made explicit. But only by making theories, conceptual
frameworks, and other conceptual constructs and presuppositions explicit
can we be rational in our consideration of the meaning and use of
empirical results in environmental toxicology. Thus, in succeeding sections,
we will deal with conceptual considerations and theoretical considerations
as well as empirical considerations of importance in the design, conduct,
and evaluation of laboratory research. And we must keep in mind that the
results of such research in environmental toxicology are intended to
reduce problems, of-toxic materials in natural systems.
But before beginning to do this, there are two other considerations
that may be helpful. We have already noted that internal consistency is
an important explanatory criterion: our arguments ought not to be self-
contradictory. But they must also be externally adequate, or they are
10
-------
m
-a
o
CD
Figure 1. The conceptual continuum of levels of explanation and understanding of an individual
scientist. This is considered to be within the context of the scientific community. After Warren,
Allen, and Ilaefner (MS).
-------
irrelevant. Mere prediction is not enough: our explanations should
account for what we know or believe to be involved in the natural systems
we would explain. And because no explanation can be final, the heuristic
power of explanations is another important criterion—their potentials
for developing further more adequate explanations. Visualizability,
another criterion, is involved in this. Aesthetic appeal may well be the
ultimate criterion, because it would surely involve the other criteria,
but it would also involve much more of our feelings about nature, which
we ought not to dismiss lightly.
The final introductory consideration is the matter of universal
explanation. Plato saw this as being possible only through abstract
general and invariant forms, which were to be taken as being ultimately
most real. This, of course, is idealism, which discounts the reality of
the ever-changing and often apparently inconsistent flux of human per-
ceptual experience. Physics, in the main, has followed Plato, with
considerable success. Biology has more nearly followed Aristotle, the
father of empiricism, who was suspicious of theoretical constructs and
believed that universal explanation was implicit in the data of human
experience. To many biologists, then, the answers to our questions—
the solutions to our problems— seem to lie in the accumulation of more
data. But this has not led to much if any universal explanation. We
sometimes fail to remember that the data are the problem, for they must
be explained, and this can be done only by conceptual constructs. More
data do not usually decrease our difficulties of understanding, at least
not directly so, because more data generally only show our previous
understanding to have been incomplete; and even more logical unification
of the whole then has to be done.
12
-------
II. CONCEPTUAL CONSIDERATIONS IN ENVIRONMENTAL TOXICOLOGY
A. CONCEPTUAL FRAMEWORKS AS A PRIORI DESCRIPTIONS OF BIOLOGICAL SYSTEMS
Logical positivism, as a philosophy of science, largely ignored the
context and origin of theoretical propositions and thus failed to account
for scientific change. In his widely-read book, "The Structure of Scientific
Revolution," Kuhn (1962) emphasized the priority of scientific "paradigms,"
within which normal science proceeds until serious difficulties are
encountered. Solution of these problems sometimes requires new theories
of drastically different form, and old rules may be broken to achieve
these theories. Scientific revolutions occur with the crumbling and
replacement of paradigms. Mainly as a historian of physics, Kuhn saw
theories as setting the paradigms. And considering the great univer-
sality of the major physical theories, this may for physics be an adequate
view. But Newton's dynamic laws and the general theory of relativity are
no ordinary theories. For about 200 years, physicists could not even
imagine a non-Newtonian world, and their work was governed accordingly.
In our opinion, biology does not yet possess this kind of universal
theory,, Thus we have no generally accepted a priori description of the
biological world to guide our theoretical and empirical search for under-
standing. And, yet, biologists certainly have some conceptual basis for the
judgments they continually must and do make in their theoretical and empirical
investigations. But so long as that which they take to be a priori remains
unarticulated in any readily accessible form, the rationality of their
judgments and the adequacy of their theoretical and empirical work can
hardly be evaluated. Gutting (1973) defined a conceptual framework as a
description that subsumes theories and other conceptual structures pertaining
to a natural domain, along with rules for employing such a description. To
do so, such a description would need to be at a higher, more comprehensive,
level of understanding, would need to be highly abstract, and would need to
entail what most fundamentally and universally underlies all phenomena in
the domain of interest. Warren, Allen, and Haefner (MS) tentatively proposed
such a conceptual framework for biology, in which they articulated abstract
concepts in a set of four natural language propositions.
The most enduring contribution of Immanuel Kant was his recognition of
the logical necessity of a priori principles and categories for human per-
ception and understanding. His principles were space and time—in which we
perceive the form of objects—and causality, by which understanding of the
interrelationships of objects is possible. Thus Alfred North Whitehead was
to think of an event—an object or a system—as being four-dimensional in
space and time. Now, whatever may be the nature of reality, the only world
we could know must conform to the possibilities of human perception and
understanding, as Kant so strongly emphasized. For us to perceive and
understand a system, it must extend in space and time, and understanding
may ultimately demand some notion of causality, even though the debates on
indeterminacy will undoubtedly continue.
13
-------
Thus, we will take a biological system, at whatever level of organiza-
tion, to be composed of an organismic system together .with its level-specific
environmental system co-extending in space and time, as we attempt to
illustrate in Figure 2, The biological system is not what we perceive in
any momentary space-time section, but rather that which persists in space
and time from some origin, through the present, to some future end. For
any biological system, its present state can be understood adequately only
in terms of its past and possible future states. This is why we think it
logically most useful to define any biological system as the entire four-
dimensional event, no matter how partially we may perceive it0 Notions of
causality will be evident in our statement and various interpretations of
the following four iDiological system generalizations that Warren, Allen,
and Haefner (MS) proposed,,
1. Any performance of any organismic system is an outcome of its
operation, which consists of the interactive performances of its subsystems,
and has functions or plays operational roles in maintenance, organization,
or replication of the organismic system and a more encompassing system,
2. The potential capacity of any organismic system predetermines all
possible sequences of realized capacities, which in turn determine all
possible performances, any occurring sequence of realized capacities depend-
ing on the environmental system through time and any occurring performances
depending upon the immediately effective environment.
'3. Anytperformance of any organismic system requires space, time, .energy,
materials, and information, which are provided in particular forms and
limited amounts by its co-extensive environmental system; potential and
realized capacities determine the forms and amounts that will permit perform-
ance and thus determine the possible environmental systems within which the
organismic system could persist.
4. Any organismic system tends to incorporate in some degree not only its
organismic subsystems but also their particular environmental systems.
We take this set of generalizations to provide a minimal a priori and
outermost description of any biological system. To accomplish this,
abstract theoretical, or primitive, terms are employed. Within limits set
by their contexts in the generalizations, these terms can be interpreted in
different ways so as to be applicable to explanation and understanding of
biological systems on different levels of organization. Thus we will only
interpret or illustrate such terms; we will.not define, them. This is a
common practice in the development of theoretical scientific deductive
systems, and it becomes even more important at the conceptual framework
level of understanding (Fig. 1), if conceptual frameworks are to subsume
entire domains of investigation.
The first generalization distinguishes between the performance of an
organismic system as a whole, the operation of that system which involves
the interactive performances of subsystems on successively lower levels of
organization, and any functions or operational roles a performance of a
14
-------
BIOLOGICAL SYSTE
Past
States
Present
State
Future
States
3
33
CO
Figure 2. A biological system viewed throughout its existence as being composed of an organismic
system together with its spatially and temporally co-extensive environmental system. After Warren,
Allen, and Haefner (MS).
-------
system may have in more encompassing systems. But the performance of any
natural system is an extremely variant phenomena and cannot be universally--
in a general and invariant way—explained in operational or functional
terms. Thus, in Generalization 2, we introduce the notion of the capacity
of an organismic system. Capacity is a very abstract notion of all possible
performances of an organismic system in all possible environments. We
cannot even conceive of a system exhibiting any performance for which it does
not have the capacity. Nor can we determine, either theoretically or
empirically, all possible performances of any organismic system in all
possible environments. Yet any possible performance of any organismic
system can be very abstractly accounted for as resulting from the capacity
and the environment of the system. Thus a general, but very initial,
explanation of performance becomes possible. We believe capacity to be the
most fundamental notion required in biological explanation and that theoretical
and empirical approaches ought to be evaluated on the basis of how well
they deal with organismic system capacity.
Now it is generally recognized that the capacities of organismic systems
change through time. Depending on the level of biological organization,
this may result from either or both system development and system evolution.,
To deal with this, we must conceive of an organismic system as having some
potential capacity that, through organization depending also on the prevail-
ing environmental system, can lead to one of many possible sequences of
realized capacities, as we attempt to-illustrate in Figure 3. From the
same initial potential capacity, the sequence of realized capacities will
differ in different environments. And for a particular realized capacity,
kinds and levels of performance will differ according to the then prevail-
ing state of the environmental system (Fig. 3). We can think of organization,
structure, and replication as being very general classes of performance, to
be interpreted differently at different levels of organization.
Generalization 3 asserts the logically necessary relativity between an
organismic system and its environmental system, which together form a
biological system. This relativity has perhaps four aspects. First, an
environmental system can be defined only in relation to a specified organ-
ismic system. Second, an organismic and an environmental system can.only
persist together; we can conceive of neither as persisting without the
other. Third, in a very important sense, the potential capacity of an
organismic system determines what sets of environmental conditions would be
necessary for its persistence, and thereby predetermines all potential
environments. And fourth, it is implicit in this generalization that the
environment of an organismic system is specific to the level of organization .
of that system.
The important matter of the level-specificity of environmental systems
can be made clear through illustration of Generalization 4, which asserts
that organismic systems tend to incorporate the environments of their
subsystems, as well as the subsystems themselves. Figure 4, among other
matters, illustrates that other subsystems are in the environment of a
particular subsystem, but, by definition, these cannot be in the environment
of the organismic system as a whole. Beyond this, Figure 4 illustrates—for
16
-------
Realized Capacities
that could have
occurred had
\differenl
\ environment al
(conditions
] existed
POTENTIAL
CAPACITY ol
-—.Performances that
'•^ \could have occurred
)had some other
environmental conditions
existed at State 2
l ORGANIZATION)
^STRUCTURE /BFORMANCES
/Occurring under
^REPLICATION/ environmental
conditions existing
at State 2
Realized Capacities
that could have occurred
at state 1 had different
environmental conditions
"V ^ \ existed
Figure 3. A diagrammatic view of an organismic system showing that from some potential capacity the
organismic system passes through a particular sequence of realized capacities jointly determined by
its potential capacity and the actually prevailing environmental system. After Warren, Allen, and
Haefner (MS).
-------
oo
:o
O
O
V .
r-
m
PERFORMANCE
OF ORGANISING
SYSTEM
ORGANIZATION
STRUCTURE
REPLICAT
Possible
teractive Performanc
Figure 4. Organismic subsystems and their level-specific environments, all of which are incorporated
by an organismic system having its own particular environment. After Warren, Allen, and Haefner
(MS).
-------
a single organizational state of an organismic system—the level-specific
capacities of organismic subsystems as well as that of the system as a
whole. The capacity of any subsystem on any organizational level determines
all possible performances of that subsystem, actually occurring performances
depending also on the state of its environmental system. Operation of the
organismic system can be seen to be the interactive performances of sub-
systems on successively lower levels of organization, such operation leading
to the performances of the organismic system as a whole. The functions of
performances of the organismic system as a whole are to be seen in terms of
the operational roles of these performances in some more encompassing
organismic system.
Biological organization, from molecular to community levels, is character-
ized by ever more encompassing systems. The individual organism is one
such level of organization, the biological population another, and the
biological community an even more encompassing level. It must, however, be
emphasized that there are important levels of biological organization between
these commonly recognized ones. And especially we must emphasize that each
organismic system on each level of organization has its own particular
capacities and thus its own possible performances. Capacities and performances
on different levels of biological organization are not the same. This has
much to do with the design, conduct, and interpretation of laboratory studies
in environmental toxicology. And we should also emphasize that only the
performance of an organismic system can be more or less directly measured.
Evaluation of the capacity and the operation of an organismic system—even
though these are extremely important in explanation and application of
results—must always be indirect and very partial.
B. ON THE NATURE OF BIOLOGICAL SYSTEMS
It must be apparent from the foregoing that we, at least, take the
set of four generalizations presented earlier to define abstractly that
which is essential to understanding of all biological systems. Our discus-
sion and illustration of these generalizations and abstract concepts may
have helped to make their intended meanings clear. But to be universally
applicable throughout the biological domain, the generalizations themselves
must remain abstract, so as to be variously and usefully interpretable as
needed. Moreover, to deal with vastly different levels of biological
organization, such a set of generalizations cannot say particular things in
particular ways about particular systems. Still, to be of value in theoreti-
cal and empirical biological investigations, these generalizations must say
what cannot be left out of our understanding of biological systems. They
are intended to be employed" in both heuristic and evaluative ways. But
those of us who have been working with them, even though we have found them
to be extremely useful in our own work, are under no illusions that they
will be easily understood and readily accepted by others.
We must now proceed to interpret these generalizations for particular
levels of biological organization that concern all of us working in
environmental toxicology. In so doing, we will probably be able to make at
19
-------
least some of their possible meanings clearer, for to deal with particular
levels of biological organization, it is not only legitimate but necessary
to interpret abstract propositions in particular ways, thereby making them
more visualizable. For now, we will interpret these generalizations only
for the individual organism, the population, and the community levels of
biological organization. In doing this, we will place particular emphasis
on the differences in spacio-temporal extents, organizations, capacities,
operations, performances, and functions of biological systems on these
three different levels of organization.
Consistent with our view that adequate explanation and understanding
of any organismic system in any state- must be based on knowledge of previous
states and possible future states (Fig. 2), we interpret an organismic
system at the individual organism level of organization to be the entire
possible life history trajectory. That is, the individual organism, as a
system, is not simply what we may observe in space at any particular time
but rather is the entire continuum of states from the zygote through repro-
ductive maturity to death. Co-extensive in space and time with any develop-
ing individual organism is an environmental system,, which changes as the
needs of the individual organism change as well as with conditions influenc-
ing environmental factor levels. An individual organism together with its
environmental system forms a biological system at the individual level of
organization, which we take to be the lowest level ecological system.
' The individual organism is composed of physiological systems: the
organismic subsystems together with their level-specific environments, all
incorporated in the individual organism and all having their own develop-
mental trajectories. This brings us, then, to the matter of organisation
of the individual organism. We are free to interpret as is useful the
diagramatic representation of an abstract organismic system shown in Figure
4. Many more than the two levels of organismic subsystems here shown are
certainly present in individual organisms, and certainly there exists much
greater complexity of subsystem interactions than that represented. What
mainly we wish to emphasize is that an individual organism incorporates not
only its organismic subsystems but their environments as well, that these
organismic subsystems have their own capacities and performances, and that
the performances are dependent both on capacities and environments. Accord-
ing to our theoretical or empirical investigational needs, we may differently
interpret the levels of organization, but we should keep in mind that, with
much overlap, there are in higher organisms certain organ system, organ,
tissue system, tissue, cell system, cell, and still lower levels of organi-
zation. Even though the organization of some species, say man, at the
individual level of organization has been much studied, not nearly all of the
relations among such subsystems are known; neither, for that matter, have
the subsystems been well defined. In defining subsystems and their inter-
relationships, we are approaching very near to causal or operational expla-
nations, and these, of course, are always "written onto" the system for
explanatory purposes, never observed. Few philosophers would disagree
with this statement.
20
-------
What are we to take to be the level-specific performances of an
individual organism as a whole? Its persistence throughout its life history
trajectory is perhaps the most inclusive performance of an individual organism.
Envelopment is necessary for this and is another very high level perform-
ance. Structure at any stage of development is also a performance of the
individual as a whole, not only in that it is a result of development but
also in that it is maintained operationally through the performances of
lower level systems. And growth and reproduction are performances of
individual organisms as wholes. At most, we measure particular performances
of whatever levels of organization.
There are different ways in which we give meaning, explain, or come to
understand such performances of individual organisms as wholes. Operational
or complex lower-level causal explanations are one way. This is done by
interrelating, as best we can, those relatively few measurements of perform-
ances of the subsystems of individual organisms we are able to make. This
must always be done by conceptual structure, not observation, because we
can never hope to measure all occurring performances of all subsystems,
much less their interactions. Functional explanation is another way in
which we give meaning to performances of an individual organism as a whole.
This is done by defining a more encompassing organismic system, say a
family or a population, of which the individual organism is a part. The
functional explanation of a given performance, say reproduction, then becomes
the operational role that the given performance plays in the organization
and persistence of the encompassing system.
But we know that had the environment of an individual organism been
.different, most of its performances then would have been different. And we
know that we cannot measure all possible performances of an individual
organism in all possible environments. Still, we are aware that the per-
sistence of an individual organism in nature depends entirely on its capacity
to perform appropriately in a changing environment. Moreover, capacity
itself changes from some original potential capacity—residing in the zygote—
through some sequence of realized capacities, according to environment and
throughout the life of the organism. In no way can such capacities be
directly and fully evaluated. But the adequacy of biological explanation
and understanding depends greatly on how well the capacity of individual
organisms and other organismic systems is evaluated theoretically and
empirically. Explanation and understanding in environmental toxicology are
not exceptions to this.
We have now to interpret the biological system generalizations, and
the figures illustrating them, at the population level of biological organi-
zation,, This will necessarily be a more difficult interpretation than that
for the individual organism level of organization. Whereas an individual
organism is more or less directly observable—in other words, individual
organism is an empirical concept—populations as wholes can in no ordinary
sense be directly observed. Though it may sometimes be used as an empirical
concept, the biological population concept is much more nearly a theoretical
one. We will define a biological population, as an organismic system, to be
a system of interbreeding organisms, more or less reproductively isolated
21
-------
from other such systems of the same species, from the origin of such a
population to its extinction as a particular system.
t
Most populations are not simply aggregates of individual organisms but
rather have their species-specific and even population-specific patterns
of organization. Largely according to species, populations may have sub-
systems of families, tribes, reproductive demes, or other sorts of organi-
zation, and Figure 4 should be interpreted as necessary. Sex classes, age
classes, size classes, and social classes are generally involved. What is
important, here, is that the capacities and performances of any population
are very much determined by the nature of its organization.
The performances of a population are specific to its level of organi-
zation, as are the performances of any defined organismic system. The
organization of a population generally changes throughout its "growth" or
"development," which are to be taken as population performances. Organization
of a population is also changed by evolution, a population level performance
resulting from natural selection from among variant forms present in the
population. The structure, or composition, of a population at any time is
also a performance. And total tissue elaboration, or production, is a
performance of populations. These and other population performances, together
with the reproduction of its individual organisms, lead to the persistence
of a population in space until its extinction. Persistence is thus very
high~level population performance concept. Populations replicate, or
"reproduce" themselves, but not in the same way as do individual organisms.
We may think of a population as replicating itself when one (in the case of
asexual reproduction) or more of its individual organisms leave the popula-
,tion and colonize a location sufficiently separated from the original for a
distinct new population to arise.
For populations to exhibit such performances, which we can sometimes
measure, they must have the capacities to do so (Fig. 3). To adequately
explain population performances, as for performances of other kinds of
organismic systems, we must evaluate population level capacities for growth,
production, evolution, and persistence as fully as we are able. But, at
the population level of organization, the evaluation of capacities becomes
much more difficult than it is at the individual level. And yet, in a very
real sense, it becomes more important to do so. For it is persistence of
natural populations, not persistence of particular individual organisms,
that most concerns man, except perhaps in his own case.
Operational explanations of population level performances are based on
the interactive performances of individual organisms and other population
subsystems. Functional explanations of population level performances are
seen in the operational roles these play in more encompassing systems.
These are of two sorts. The biological species as a system of populations,
and the biological community, an altogether different sort of organismic
system within which populations are involved.
22
-------
We think it most useful to define a biological community not as a single,
relatively short-lived, extant state but rather as the entire series of
states from the origin of the community at some time in the past to some
time in the future when it ceases to exist as a particular community. Over
this rather long period of time, such a community develops from a "pioneer"
serai community, through intermediate serai communities, to some sort of
"climax" community, which is more persistent than the states or stages
preceding it. This is the community as the sere, the entire developmental
or successional sequence of serai and climax community states, which is
another possible interpretation of the abstract representations of an
organismic system given in Figures 2 and 3. We will think of such an object
as an organismic system at the community level of organization. In no way
do we wish to suggest that a biological community is any sort of an organism
or superorganism. Rather, we view it and its co-extensive environment as a
biological system definable by interpretation of the set of four biological
system generalizations we advanced.
Two views of the nature of plant communities have dominated thinking
of plant ecologists since early in this century. Unfortunately, these*
views have usually been taken to be mutually exclusive rather than to
be representative of different aspects of the same natural systems. The
organismic view of Frederic Clements (1916) and the individualistic view
of Henry Gleason (1917, 1926) and L. G. Ramensky (1924) led to a largely
fruitless controversy,, Animal ecologists generally appear to have been
either unaware of or disinterested in this controversy. Most of them seem
to have accepted the existence of biological communities as concrete natural
systems, which together with their physico-chemical habitats constitute
ecosystems.
To Clements (1916) the community, metaphorically, was an organism
"which arises, grows, matures, and dies.. „ <> able to reproduce with essential
fidelity the stages of its development." Here was the germ of one of the
truly important theoretical concepts in all ecological thought. But many
plant ecologists were distracted by Clements' logically improper use of
analogy, his too direct application of this theoretical view to vast and
sometimes fundamentally different organizations of extant vegetation, and
his treating all differences in vegetation in physically different habitats
as simply different manifestations of the same natural community to be
dealt with by a burdensome system of classification. There ensued a failure
to appreciate the best in Clements' thinking: a view of the community as a
developing system necessary for adequate explanation.
To Gleason (1917), each plant community was fundamentally individualistic,
not having the sort of genetic continuity that gives so much meaning to
taxonomic classification of individual organisms, even though communities
may be classified as a matter of convenience. He was more concerned with
how individual organisms and species, according to their particular require-
ments, become involved in the development of what we take to be communities.
Because each' location on the earth is physically in some degree different
from every other, because organisms of each kind are quite specific in
their requirements, and- because the opportunity for organisms to colonize
23
-------
different locations is not always the same, no two communities can develop
in precisely the same way. Moreover, communities intergrade one with
another, in such a way that their physical delimitation in space is generally
difficult. Their definition as objects must be of a different sort than we
employ in defining individual organisms. Gleason's view thus addressed an
aspect of plant communities not accounted for in the view of Clements.
Envisionment of the nature of natural systems that we might come to better
understand them demands as full an account of their different aspects as we
are capable of achieving at any given time. We must somehow make the best
possible use of the insights of these two great plant ecologists, and all
other relevant knowledge.
Failure to achieve a useful synthesis of these views of different
aspects of communities resulted mainly from insistence by some that there
be empirical demonstration of what communities are in reality and from
failure to recognize that any explanation whatsoever presupposes at least
some a priori conditions of understanding—ultimately the conditions of the
possibility of knowing. Scientists implicitly or explicitly must make
judgments related to the metaphysical problems of reality. But no philosopher
'or scientist has shown the way to solution of this problem since David Hume
and Immanuel Kant raised the problem of metaphysics into high relief,
Clements' view was of a plant community—the whole—as an object or system
having four dimensions in space and time and capable of developmental
performance and persistence. Gleason was more concerned with how such
community level performances—development and persistence—resulted from
the interactive performances of community subsystems, individual organisms
and species populations, which are involved in the operation of the community
as a whole and thus in its causal explanation. No view of community organi-
zation adequate for explanation can disregard either of these aspects,
.which lie at the.very basis of the possibility of our knowing about communi-
ties. We must, then, direct our attention to what we may be willing to
take as an a priori, outermost definition of what any biological community
would have to be like for us to perceive and understand it, if we are to be
able to develop a rational basis for the design and evaluation of laboratory
ecosystem studies as an aid to understanding natural communities and problems
of environmental toxicology.
As already noted, we take the four biological system generalizations
"we have advanced to be interpretable -at the biological community level of
organization. They thus constitute an a priori theoretioal view of the
fundamental nature of communities in relation to their environmental systems.
But any such theoretical view of the nature of communities is of value only
if it is reasonably interpretable in terms of what we take to be natural
communities and their parts. Beyond this, it is the function of theoretical
propositions to order observational experience so as to make it comprehensible,
not merely to reflect observational experience as do empirical generalizations.
Fundamentally, the problem of interpretation is the recognition of natural
objects that could be expected to conform to the theoretical propositions.
At any time, given what we take to be a community, what are we to take to
be its subsystems? All too often, it seems to us, ecologists have tended
to suppose that communities are simply the level of organization immediately
24
-------
above species populations, that communities can somehow be explained without
taking into account intermediate levels of organization that may exist,
But what might be these intermediate levels of organization, these subsystems
lying between the community and its individual species populations?
This is not simply an empirical question, because how we choose to
define such intermediate level subsystems will in large part determine
observational experience and how we come to understand it. In any natural
community, some populations interact more closely than do others. The
individuals of some populations consume one another, compete for common
resources, and may mutually favor one another. Such closely interacting
populations can perhaps be usefully viewed as trophic subsystems, to be
defined as a level above that of populations, in interpreting the abstract
representation given in Figure 4. But surely some trophic subsystems inter-
act more closely than do others, and this ought to be taken into account in
any theoretical view adequate for community explanation. Trophic subsystems
associated with trees and relationally involved plants and animals must
interact more closely with one another than they do with the trophic subsystems
associated with the shrubs, the herbs, or the substrate. Thus, we may be
able to usefully think of a tree subcommunity} a shrub subcommunity3 a
herb-grass subcommunity3 and a substrate subcommunity, together forming the
community as a whole and each having its own trophic subsystems. In all
this, there would necessarily be broad overlap of community subsystems at
all levels, or we could not suppose the community as a whole to be a natural
system.
Conceptualization and measurement of performances of organismic
systems so extended in space and time and so diffuse in organization as are
communities can be no simple matter. Community persistence is probably the
most encompassing performance of an individual community,, Community
development or succession, through a series of successional communities or
stages, is perhaps the most generally recognized community-level performance.
And community structure at any time, just as change in structure with develop-
ment, is a performance, and is the one that is most generally measured.
Communities maintain individual organisms of many different species popu-
lations; and these sometimes leave the boundaries of a given community to
become, through colonization, involved in the establishment of another
community. This may perhaps be thought of as community replication. But
such replication of a community is not precise, for the colonizers of
communities may come from communities of different kinds. Communities have
the function of maintaining the fauna and flora of a region, these being
necessary to the establishment and persistence of the different communities
in any region.
We reach a very high level of abstraction when we think of the potential
and realized capacities of systems so abstract themselves as are biological
communities. And yet, if communities have performances such as persistence,
development, and structure, which they most certainly do—or we could not
know of them—then it is logically necessary that communities have the
capacities for these performances. In some sense, the capacities of communi-
ties are based on the flora and fauna of a region and on all possible inter-
active performances of the species of these.
25
-------
C. ON THE ROLES OF CONCEPTUAL FRAMEWORKS IN ENVIRONMENTAL TOXICOLOGY
To many, if not most, of our readers interested in environmental toxi-
cology, this lengthy and involved discussion of conceptual frame-
works and the nature of biological systems may seem quite irrelevant to
their work in toxicology. In all probability, our readers will not before
have come upon a discussion quite like this, for we are unaware of any
similar ones, even though we have searched., And, in all sincerity, we
must confess that were we, for the first time, to have come upon such a
burdensome presentation, we would be uncertain as to what it all meant—
what role it might play in our thinking and work in environmental toxicology.
But about 25 years of work in physiology, ecology, and environmental toxi-
cology at our laboratory has convinced us of the need for explication of
the conceptual frameworks within which.we proceed in our theoretical and
empirical investigations, Finding that others had not done so, and also
that their work as well as ours exhibited weaknesses owing in large
part to conceptual considerations, we were compelled to do what we could
at the conceptual framework level of understanding, even in the knowledge
that what we could do would be limited and surely require much continued
effort on the part of ourselves as well as others. We now owe it to our
readers to emphasize the necessity and roles we see for the sort of con-
ceptual considerations with which we have been working.
In the introductory section of this report, we attempted to sketch
the complexity of problems in environmental toxicology and of the social
and natural systems within which these problems are perceived to occur.
And there we emphasized that our theoretical and empirical investigations
in environmental toxicology must relate in some way to these problems and
natural systems, if there is to be much hope of resolving the rather
serious toxicological problems we face. But how are we to evaluate our
work, to identify and implement necessary improvements in our efforts to
resolve these problems, and to recognize the limits or extent of applica-
bility of our results to natural systems? In general we believe that
adequately articulated conceptual frameworks have important heuristic,,
evaluative, and application voles to play in the conduct and employment
of environmental toxicology.
To be effective in these roles, a conceptual framework must provide
an adequate, a priori, outermost description of all biological systems,
for only in so doing can it provide a conceptual context of theoretical
and empirical work and its application. It must somehow be more universal—
more general and invariant—than are the theories and results of empirical
work it encompasses. This can be accomplished only with very high level
theoretical propositions that in some abstract way include characteristics
of biological systems that theories and empirical results either cannot or
do not include. But the abstract concepts must be interpretable at par-
ticular levels of organization in particular ways, or they would be
irrelevant to theoretical and empirical investigation and to what we
take to be natural systems. The abstract biological system generaliza-
tions Warren, Allen, and Haefner (MS) tentatively proposed—in their
considerable abstract universality and their various possible interpretations
26
-------
for different kinds and levels of biological systems—make possible much
logical unification of our knowledge.
In providing a more comprehensive view of the nature of biological
systems, a conceptual framework helps us to see more clearly the most
important problems to be solved. Moreover, it helps to suggest theoretical
forms and the designs of empirical investigations that are promising for
our work in environmental toxicology. These are important heuristic roles.
A conceptual framework, insofar as it is adequate, should more nearly
represent what is most important about biological systems and thus provide
a basis for evaluating our theoretical and empirical work. At a minimum,
because the framework is explicated, different pieces of scientific investi-
gation can be rationally considered, even if there be disagreement about
the validity of the a priori propositions. Such disagreement can rationally
be resolved only by seeking out assumptions about nature even prior to
these propositions. The rationality of science has suffered whenever
there has been failure to do this. We believe that much theoretical and
empirical work in biology can be evaluated in the light of the four
biological system generalizations. Evaluation can be made more objective,
within the conceptual boundaries of a framework, if its rules are made
explicit. This is rarely done, even though rules of some sort are surely
generally operative in science. Warren, Allen, and Haefner (MS) did
derive from their conceptual framework at least some of the rules that
seem to be implicit in it:
1. Only the performance of an organismic system or subsystem can "be
measured, its capacity and its operation "being representable only indirectly
and incompletely.
2. Measurements of the performance of an organismic system or subsystem
without relevant measurements on its co-extensive, level-specific environ-
mental system are of little explanatory value.
3. Operational explanation of an organismic system should take into
account the performances and operations of subsystems on successively
lower levels of organization and cannot be based only on knowledge .of
subsystems on the lowest levels of organization.
4. Explanatory generalizations pertaining directly to any one level of
organization of an organismic system should contain at least one concept
specific to that level and should subsume conceptual, methodological, or
other-sorts of indeterminacy that may exist with respect to lower levels
of organization.
5. Perception and explanation of organismic and environmental systems
are always partial relative to the space and time dimensions and the
components of the systems.
Many biologists will find more acceptable the generalizations earlier
reviewed than these ,rules we believe to be important, even though the
27
-------
rules appear to emerge quite directly from the generalizationsc This is
an inevitable consequence of making personal conceptual frameworks explicit.
However this may prove to be, the rationality of judgments based on a priori
generalizations and rules made explicit can at least be evaluated by others.
The roots of any dissatisfaction with such rules should be sought in
the generalizations from which the rules appear to us to follow quite
directly,, However others may feel about these particular generalizations
and rules, we strongly believe that more effort of this sort is necessary
in environmental toxicology.
There is at least one other kind of role articulated conceptual
frameworks should be able to play in environmental toxicology: that is,
as an aid in application of theoretical and empirical results to the solu-
tion of problems of toxic substances in our environment. In being at a
higher conceptual level, such frameworks give meaning to theoretical and
empirical explanations. Part of such meaning is in identification of
boundary conditions, which must be taken into account in application of
results to solution of problems occurring in natural systems. But there
is also a positive aspect of further meaning we can give to theoretical
and empirical work: it helps us to see ways of applying our results we
might otherwise not have perceived., In later sections of this article, we
will illustrate these heuristic, evaluative, and application roles of
conceptual frameworks in theoretical and empirical investigation of
environmental toxicology.
28
-------
III. THEORETICAL CONSIDERATIONS IN ENVIRONMENTAL TOXICOLOGY
A. ON THE NATURE AND ROLES OF THEORIES AND MODELS
We have now to introduce theoretical considerations we believe to be
of some importance in dealing with problems of toxic substances in man's
environment. In doing so, we wish to distinguish clearly between what we
have been calling conceptual frameworks and what we take theories to be.
The generalizations of the conceptual framework do not themselves constitute
a theory, and they are not intended to, because we envision them as operating
at a higher level of explanation and understanding (Fig. 1). Theories, to
accomplish their purpose, must make it possible for us to say more particular
things about more precisely defined systems. We must now make clear what we,
at least, take to be the nature and roles of theories and models.
For, unlike in physics, there is in biology no general agreement as to what
a theory may be, the term theory and the term model often being employed
vaguely and even interchangeably.
Although it is not at all clear to us that adequate biological theory
can have the same form as physical theory, we will in general restrict^our_
meaning of theory to that employed in physics. Here, a theory is a scientific
deductive system beginning with a set of initial propositions, at least one
of which is a theoretical proposition by virtue of its involving a theoretical
concept (Braithwaite, 1953). From the initial propositions, theorems are
deduced by means of a calculus. But because the calculus itself adds no
new meaning, any such theorems remain abstract, in that they still contain
the undefined theoretical concept. For the theorems to be related to the
perceptual world, rules of correspondence are employed to translate the
theoretical concept into one or more empirical concepts—of phenomena more or
less directly observable. Thus the theorems are translated into empirical
generalizations, which are related only indirectly to the initial theoretical
generalizations via the rules of correspondence and back through the abstract
argument of the calculus.
In this way, empirical generalizations are explained by their deduction
from theoretical generalizations. A number of different theorems may be
deducible from the same set of theoretical propositions. These theorems, and
any empirical generalizations into which they may legitimately be translated,
are thus logically unified. Moreover, because the same theoretical concept
can be translated, or interpreted, in different ways, a wider range of human
perceptual experience can be explained. This is the primary function of any
theory. But these very characteristics of theory, so understood, give it
heuristic potential. Gpod theories are quite universal: they explain
broad ranges of human perceptual experience in general and invariant
ways.
This is something no empirical generalization can do. For all that
can be deduced from and so explained by an empirical generalization are
the same, or very nearly the same, observational experiences from which it
was originally induced. Thus, without theory, we can know little about
the meaning and generality of an empirical generalization. And therefore,
29
-------
say in environmental toxicology, we are unable to know with any certainty
at all how generally an empirical generalization can be applied in the
solution of problems. We can, of course, extend our knowledge of its
'generality and applicability by further empirical investigation. But
there are so many differences in natural systems and the conditions under
which they exist—and thus in the performances we might possibly measure—
that we cannot hope to deal adequately with problems in environmental
toxicology by empirical means alone. Moreover, were we to accumulate
large amounts of data in attempting to determine the generality of some
particular empirical generalization, we would usually find it to be very
limited, and perhaps many different empirical generalizations to be
necessary to describe these data. The reasons for this—or the relation-
ships among these empirical generalizations and fundamental characteristics
of any natural system of interest—would be unlikely to be obvious. But
were we to be able to conceive of a set of theoretical propositions and a
deductive argument making the empirical generalizations logical consequences,
then we would have more adequately explained as well as logically unified
not just the generalizations but the natural system itself.
There are very good reasons for distinguishing between models and
theories. Models are made more particular and visualizable than theories,
usually by empirical concepts, by analogy to observable systems—thinking
of atoms as little solar systems—or by construction of physical represen-
tations. Theories are difficult to think about because they are abstract,
and models of theories are sometimes employed, since in their particularity
and visualizability the models help to alleviate this difficulty. But in
so doing, models lose much of the universality of theories. Mathematical
models are models rather than theories because their symbols are interpreted
a priori in terms of observable events. Laboratory ecosystems are models
of natural ecosystems, so long as our interest is in natural ecosystems,
not simply in what happens in the systems we have created in the laboratory.
They are models because they are representations composed of visualizable,
measurable components. Such models make possible the generation of data
that may be relevant to explanation of natural ecosystems. In this they
share a characteristic with computer simulation models. Of course, the
relevance of any such data to explanation of natural ecosystems depends
very much on the representational adequacy of the model employed.
Judgments as to the representational adequacy of a model ought to
take into account not only the objective the model was designed to accomplish
but also all relevant theoretical and empirical knowledge of the natural
system being modeled. If we are so fortunate as to have a formally articu-
lated and rather universal theory of such a system, then the relationship
between the theory and one sort of model is quite straightforward. By
interpreting the abstract concepts in the initial propositions of.the
theory a priori—rather than a posteriori as is done with rules of
correspondence in the theory—then the same calculus can be employed in
both the model and the theory (Braithwaite, 1953). This mode of inter-
pretation results in a more visualizable deductive explanatory system, as
the model is, but it leads to loss of explanatory and heuristic power.
30
-------
But even theories are partial in that they are apt to leave out something
if not much of what we may know about natural systems of interest. Some
of this other knowledge may be quite relevant to judgments about repre-
sentational adequacy of laboratory ecosystems0 It is for this reason that
we placed so much emphasis on a conceptual framework and the apparent
nature of biological systems. In any event, we must emphasize that models—
whether they are of explanatory form or are employed only as "tools" of
explanation—are almost always quite particular. What we can conclude
from them is not likely to be very general: universal explanation and
understanding require theories and conceptual frameworks.
With respect to theories, not all philosophers agree that they are
explanations in any ordinary sense, that is, statements that might be
characterized as true or false, good or poor. Some philosophers hold an
"instrumentalist" view: theories are primarily logical instruments "for
organizing our experience and for ordering experimental laws" (Nagel,
1961). Our own position, in what is really a continuum of beliefs as to
the cognitive status of theories, is that they serve both functions
'remarkably well: universal theories are not cnly profound explanations,
they are also powerful tools for explanation. Models that in some way
translate what was unfamiliar into what is familiar and satisfy the need
for explanation are also explanations. But, as we have already noted,
models are more particular explanations; they do not have the univer-
sality of theories, even of the theories they may interpret. And models
having some forms are not explanations at all. Laboratory ecosystems
are not explanations, because data generated with such systems still need
to be explained theoretically or otherwise. Laboratory ecosystems and
other models sharing this characteristic are tools of explanation, ways of
finding explanation, and they should be used in conjunction with theories
and other explanatory systems whenever possible.
As to the form, function, and efficacy of models, we have
noted two very different functions of models: an explanatory function
and an instrumental or "tool-like" function. The form of a model very
much determines whether it can be efficacious as an explanation, as a tool
for explanation, as both, or as neither. Not only the evaluation but also
the design of models ought to be based on their intended functions or
objectives, for a model of a given form is not apt to serve several functions
well, if at all.
Now models have many uses, some as explanations, some as aids to
explanation, and some in making more apparent those phenomena that require
explanation. Because models make the abstract notions of theories and
other conceptual structures more concrete, visualizable, and manipulatable,
they may serve as aids in the analysis of conceptual structures and the
phenomena these represent. It is sometimes claimed that models are useful
in further development of an extant theory. Braithwaite (1960), in a
careful evaluation of the capacities of models, rejects some such claims
but concurs that models may be of genuine assistance in theory development
if they add hypotheses further relating the theoretical concepts of the
theory or if they add hypotheses with new theoretical concepts interpreted
31
-------
in familiar ways. A model may suggest similarities in theories previously
considered to be distinct and thus may show the way to bring two such
theories together under a higher level theory; at least this is a claimed
objective of modeling in general systems theory (von Bertalanffy, 1968).
Beyond all this, models may be employed to outline problems and guide
research effort, to generate new data, and to make predictions not always
simply derivable from theories and other more abstract conceptual structures,
Finally, we should not attribute to the theory characteristics of the
model owing simply to its modality (Braithwaite, 1953; Nagel, 1961; Toulmin,
1960); yet, in our thinking about natural phenomena, we should never
uncouple a model from the theory or higher level conceptual constructs it
may only poorly represent.
Having considered the importance of keeping in mind the higher level
conceptual structures from which models derive, let us return to the impor-
tance of designing and evaluating models according to their specifically
intended functions, whether these be of an explanatory or of an instrumental
nature. For carpentry, hammers are designed for pounding, screwdrivers
for setting screws, saws for cutting boards, and planes for smoothing. A
carpenter does not use a hammer for setting a screw nor a saw for smoothing
a board. But neither does he use a ball-peen hammer for driving nails nor
a large-toothed saw for finished cabinet work. And so it is with models,
either as tools or as explanations. In the design and evaluation of
laboratory ecological systems as models, our objectives—their intended
functions—should be borne in mind, as should the conceptual structures by
which we intend to give meaning to any derived results, or to relate these
results to natural systems.
32
-------
Bo THE DYNAMICS OF ORGANISMIC SYSTEM PERFORMANCE
One of the functions of biological theories and models is to begin to
explicate the effects of environmental factors on the performances of
organismic systems. In this section, we will illustrate how a deductive
argument developed by Booty and Warren (MS) can be used to examine some
of the possible effects of changes in the levels of environmental
factors on both dynamic and steady-state performances of simple multispecies
systems.
In Figure 5A, a simple biological system is represented as a sequence
of predator-prey interactions. The defined system is composed of inter-
acting carnivore (C), herbivore (H) and plant (P) populations, and light
resource level (R). Fishing effort (E) and rate of light input
(I) constitute the environment of the defined system.
Phase planes and interrelated isocline systems representing this
sequence of predator-prey interactions are shown in Figure 6. On each
phase plane, prey biomass (or resource level) is plotted on the x-axis and
predator biomass (or utilizer level) on the y-axis. The descending ourves
parameterized by light input rate (low I, med I, high I) on each phase
plane.are the prey or resource isoclines. Each prey isocline is a set of
biomasses (or levels) of predator and prey at which the rate of change of
prey biomass with time is zero (dR/dt = 0 on the PR phase plane, dP/dt = 0
on the H-P phase plane, and dH/dt = 0 on the C-H phase plane). The
ascending curves on each phase plane are predator or utilizer isoclines.
Each predator isocline is a set of biomasses (or levels) of predator and
prey, at which the rate of change of predator biomass with time is zero
(dP/dt = 0 on the P-R phase plane, dH/dt = 0 on the H-P phase plane, and
dC/dt = 0 on the C-H phase plane). Each predator isocline is parameter-
ized by a particular herbivore biomass on the P-R phase plane, by a
particular carnivore biomass on the H-P phase plane, and by a particular
level of fishing effort on the C-H phase plane. Each intersection of a
predator and prey isocline, where the rate of change of both predator and
prey biomass with time is zero, is a steady state point.
The forms, positions, and identities of all isoclines on the phase
planes in Figure 6 have been graphically deduced from response functions
that represent biological characteristics or performances of each of the
interacting populations. Any population performance is determined by the
population's capacity and its immediately effective environment, which
includes other populations of predators, prey, and competitors. A deduc-
tive system, by logically unifying interactive population performances,
can explicate performances of the higher level system of which the
populations are parts. Such a higher level system, as a whole, is shown
in Figure 5A. With the phase planes and isocline systems shown in Figure
6, the effects of changes in the level of environmental factors I and E on
one of the performances of the defined system, its steady-state structure,
can be determined. First, however, we will illustrate the graphical*
method of deducing predator and prey isoclines.
33
-------
A.
E
-It
C
H
11
P
It
R
-+
-*
C.
H^H,
l+
-J
R
Figure 5. Kinetic diagrams representing defined multispecies system, including plants (P) , herbivores
(H), and carnivores (C). Rate of light input (I) and rate of exploitation (E) represent external
environmental variables. Light intensity (R) is a variable dependent on I and P. Competition,
commensalism, and mutualism are represented in B. And a toxicant (T) directly affecting only carnivore
C is represented in C.
-------
The form and position of the isoclines on each phase plane can be
deduced from systems of curves representing such population performances
as recruitment, production, loss to predation, nonpredatory losses, and
yield to exploitation. Although there is a considerable body of theoretical
and empirical knowledge of the forms of such curves or response functions,
we cannot expect to determine them empirically •under the conditions
defining them. Rather they are theoretical relations allowing us to take
into account much of what we do know or wish to hypothesize. Booty and
Warren (MS) provide a detailed description of the graphical derivation of
predator and prey isoclines on all phase planes. Here, only derivation of
the predator and prey isoclines on the C-H.phase plane will be shown. The
same general procedure is used to determine the forms and positions of the
isoclines on the P-R and H-P phase planes.
»
The position and form of the prey isoclines on the C-H phase plane
are derived from response functions that represent performances of the
herbivore population, as we show in Figure 7. The rate of change of
herbivore biomass with time (dH/dt) is dependent upon the rate that the
herbivore population gains biomass and the rate that it loses biomass.
The population of herbivores in this simple community gains biomass through
recruitment and production and loses biomass through consumption by carni-
vores (predation losses) and nonpredatory losses such as those owing to
emigration and disease.
Herbivore recruitment and production gain curves are derived from
the prey isoclines on the H-P phase plane and curves relating herbivore
relative recruitment rate and relative growth rate to plant biomass.
Derivation of herbivore production gain curves is illustrated in Figure
8. The prey isoclines (dP/dt = 0) on the H-P phase plane define density-
dependent relationships between herbivore and plant biomasses at
different rates of light input. A herbivore production gain curve and
a herbivore recruitment gain curve can be determined from each prey
isocline. Each production and recruitment gain curve is, therefore,
parameterized by a particular light input rate. Herbivore biomass is
removed from the herbivore population through consumption of herbivores by
carnivores (predation losses, Fig. 7E) and nonpredatory losses (Fig. 7F).
The recruitment and production gain curves of the herbivore for each rate
of light input are summed to generate a set of curves representing the
total gain rates of the herbivore (Fig. 7D). Similarly, the nonpredatory
loss curve is summed with each predation loss curve to generate a set of
curves representing herbivore total loss rates (Fig. 7G).
The prey isoclines on the C-H phase plane are defined as the sets of
biomasses of carnivore and herbivore at which dH/dt = 0. When dH/dt = 0,
the total rate of gain of herbivore biomass is equal to the total rate of
loss of herbivore biomass. If curves representing total gain rates (Fig.
7D) and curves representing total loss rates (Fig. 7G) are plotted on the
same graph, as shown in Figure 7A, the intersections of a gain curve
identified by a particular rate of light input with a series of loss curves
identified by carnivore biomasses define a set of biomasses of carnivore
35
-------
Figure 6. Phase planes and interrelated isocline systems representing a
sequence of predator-prey interactions in a simple community are shown in A,
B, and Ce The form and position of predator and prey isoclines on all phase
planes have been deduced, with a graphical calculus, from response functions
that represent the biological characteristics of each of the populations.
An infinite family of prey isoclines exists on each phase plane. Each prey
isocline is generated and so parameterized by a particular rate of light
input I. The prey isoclines generated by three rates of light input, low I,
med I, and hjLgh I, are shown on each phase plane. An infinite family of
predator isoclines also exist on each phase plane. Predator isoclines are
generated and thus parameterized by particular herbivore biomasses on the
P-R phase plane; by particular carnivore biomasses on the H-P phase plane;
and by particular levels of fishing effort on the C-H phase plane. The
solid circles on each phase plane are steady-state points defining the
steady-state biomasses of C, H, P, and R at high I, OE. The solid squares
define the steady-state biomasses of C, H, P, and R at med I, 90 E.
36
-------
5LOW MEOI HIGH I
LOW I HED I HIGH I
LOW I MED I HIGH 1
4-
Figure 6.
-------
Figure 7. Derivation of prey isoclines on the C-H phase plane (Figure
6C) by graphical summation of herbivore gain and loss response functions.
The prey isoclines on the C-H phase plane are the sets of biomasses of herbi-
vore and carnivore where dH/dt = 0. The intersections of the curve represent-
ing herbivore total gain rates for a particular rate of light input I, with
the series of curves representing herbivore total loss rates for different
carnivore biomasses, C, (A) defines the set of herbivore and carnivore
biomasses where dH/dt = 0. Each rate of light input generates and thus
parameterizes a particular prey isocline. The curves representing the total
gain rates of the herbivore (D) are constructed by graphical summation of
herbivore recruitment (B) and production (C) rate curves for each rate of
light input. The curves representing herbivore total loss rates (G) are
constructed by graphical summation of the nonpredatory loss rate curve (solid
line, F) with each predation loss rate curve (E). Total loss rates at OC
represent only nonpredatory losses. The nonpredatory loss rate response
could also be represented by a family of curves parameterized by different
levels of chemical, physical, or biological factors that would affect non-
predatory loss rate (dashed lines, F).
38
-------
GAIN
LOSS
5C 5C 1C OC
<£>
o
-H
SO
o
o
o
CO
\-
z
UJ
S
5
-------
Figure 8. Derivation of the herbivore production gain response functions
used in constructing the prey isoclines on the C-H phase plane. Herbivore
production rate is the product of herbivore relative growth rate and
herbivore biomass. Each production gain curve is derived from the density-
dependent relationship between herbivore and plant biomass at a
given light input rate and from the curve relating herbivore relative growth
rate to plant biomass. The determination of herbivore production rates from
these relationships is illustrated for herbivore biomasses 1 and 3 at light
input rate med I. In general, at a given rate of light input, an increase
in carnivore biomass results in a decline in herbivore biomass and an
increase in plant biomass. With an increase in plant biomass, the relative
growth rate of the herbivore (gi) also increases. Herbivore relative growth
rate is therefore an inverse function of herbivore biomass and the relation-
ship between herbivore production rate and herbiv ore biomass is a dome-shaped
curve. Each rate of light input defines a unique density-dependent reltion-
ship between herbivore and plant biomass and, thus, a unique herbivore
production curve.
40
-------
LOW I MED I HIGH I
H
UJ
CE
o
Q
O
cr
Q.
oc
5C
UJ
CE
X
I
o:
° *2
UJ
>
H 9
-------
and herbivore at which dH/dt = 0. That is, each rate of light input
generates and so identifies a prey isocline on the C-H phase plane ^shown
in Figure 6C.
In general, this procedure is used to derive prey isoclines on all
phase planes. Rates of light input are initially introduced as the gain
terms of R and so become the identities of prey isoclines on the P-R phase
plane shown in Figure 6A. Rates of light input also become identities of
prey isoclines on all successive phase planes. This is a consequence of
constructing prey isoclines on any particular phase plane by utilizing
recruitment and production gain curves derived from the prey isoclines of
the previous plane.
Rate of light input is not the only possible parameterizing identity
of prey isoclines. Any biological, physical, or chemical factor, includ-
ing toxicants, affecting the response functions of a population would
become an additional parameterizing identity of the prey isoclines derived
from that population's response functions.
Essentially the same procedure is used to construct predator isoclines
on the C-H phase plane (Pig. 6C): graphical summation of carnivore gain
curves (recruitment, production) and loss curves (yield, nonpredatory
losses), as shown in Figure 9. Carnivore production and recruitment response
functions are solely derived, respectively, from a relationship between
carnivore relative growth rate and herbivore biomass, as shown in Figure
10, and a relationship between carnivore relative recruitment rate and
herbivore biomass. This method of determining carnivore gain response
functions facilitates construction of predator isoclines.
Carnivore loss curves are represented by yield response functions,
parameterized by level of fishing effort (Fig. 9E) and nonpredatory losses
(Fig. 9F). Carnivore recruitment and production gain curves at each
herbivore biomass are summed to generate a set of curves representing
carnivore total gain rates (Fig. 9D). Similarly, the nonpredatory loss
curve is summed with each yield curve to generate a set of curves repre-
senting total loss rates of the carnivore (Fig. 9G). The curves representing
total rate of gain (Fig. 9D) and the curves representing total rate of
loss (Fig. 9G) can now be plotted on the same graph, as shown in Figure
9A. The intersections of a loss curve (identified by a particular level
of fishing effort) with a series of gain curves (identified by herbivore
biomasses) define a set of herbivore and carnivore biomasses at which
dC/dt = 0. -Thus each level of fishing effort generates and therefore
identifies a predator isocline on the C-H phase plane shown in 6C.
Having illustrated the deductive procedure used to determine the
position and form of the isoclines on the phase planes in Figure-6, we can
now examine the effects of changes in the levels of environmental
factors I and E on the steady-state biomass performances of populations
composing the defined system, and thus possible steady-state structures of
the defined system shown in Figure 5A.
42
-------
We can begin, on the C-H phase plane in Figure 6C, with the steady-
state point at the intersection of the prey isocline parameterized by high I
and the predator isocline parameterized by OE. Progressively higher levels
of fishing effort (OE to 150E) at this light input rate lead to lower
steady-state C biomasses and higher steady-state H biomasses. In effect,
then, increasing fishing effort moves steady-state biomasses downward on the
prey isocline identified by high I, this reducing steady-state C biomass
and, thus, allowing the steady-state biomass of H, the prey, to increase.
Now, on the H-P phase plane at high I (Fig. 6B), an increase in steady-state
H biomass leads to a reduction in the steady-state biomass of P, the steady-
state values shifting up along the high I prey isocline with declining C
biomass. And, finally, in Figure 6A, the reduction in steady-state P
biomass resulting from increased H biomass leads to an increase in R.
Similarly, at a given level of fishing effort such as 90E (Fig. 6C),
progressively higher rates of light input (low I to high I) lead to higher
steady-state biomasses of both C and H, the location of the steady-state
point now moving upward along the predator isocline identified by 90E.
Increased I causes not only increases in C and H biomasses but also increases
in P and R. Thus, changes in either I or E result, either directly or
indirectly, in changes in the steady-state biomasses of all populations
in this simple community.
There exists, then, for a given rate of light input and level of
fishing effort, a single steady state on each phase plane, the set of
these defining mutual steady-state biomasses of C, H, P, and R. In Figure 6,
the steady-state points on each phase plane that define the mutual steady-
state biomasses of C, H, P, and R~the steady-state structure of the
defined system—at high I, OE (solid circles) and med I, 90E (solid squares)
are shown. Each of the three steady-state points in a set (either circles
or squares) generated by a particular rate of light input and level of
fishing effort is simply a two-dimensional projection of a four-dimensional
community steady-state point. At high T, OE, the populations composing
the defined system attain steady-state biomasses of 4.3C, 2.5H, 3.5P, and
2R. Similarly, at med I, OE, the steady-state biomasses of these populations
are 2C, 3H, 2.2P, and 1.7R. Because there are conceivably an infinite
number of particular rates of light input and levels of fishing effort, an
infinite number of community steady-state points and, thus, steady-state
community structures exists, each generated by a particular combination of
light input rate and level of fishing effort.
In the foregoing discussion, only predator-prey interactions in a
defined system have been considered. Somewhat greater dimensionality can
be introduced by including, in the defined system, not only predation but
also competition and commensalistic and mutualistic interactions between
populations, as shown in Figure SB (Booty and Warren, MS; Liss and Warren,
MS). Biomasses of competitor, commensal, or mutualist become identities
of prey and predator isoclines in addition to light input rate and level
of fishing effort and, thus, also determine steady-state population bio-
masses, population persistence, and community structure.
43
-------
Figure 90 Derivation of predator isoclines on the C-H phase plane (Figure
6C) by graphical summation of carnivore gain and loss curves. The predator
isoclines on the C-H phase plane are the sets of biomasses of carnivore and
herbivore where dC/dt = 0. The intersections of the curve representing
carnivore total loss rates for a particular level of fishing effort, E,
with the series of curves representing carnivore total gain rates at dif-
ferent herbivore biomasses, H, (A) define the set of carnivore and herbi-
vore biomasses where dC/dt = 0. Each level of fishing effort generates and
so identifies a particular predator isocline. The curves representing the
total gain rates of the carnivore (D) are constructed by graphical summation
of carnivore recruitment (B) and production (C) rate curves for each herbi-
vore biomass. The curves representing carnivore total loss rates (G) are
constructed by graphical summation of the nonpredatory loss rate curve (F)
with each yield rate curve (E). The problematical yield response curves are
determined from dY/dt = qEC where q, the catchabi.lity coefficient, was fixed
at 0.2. Total loss rates at OE represent only nonpredatory losses.
44
-------
3J.VU NIV9
31VH SSCTI 1V10J. 0
3iVH SSCH 1V.LOJ. 0
V X »
3iva
31Vd SS01 Aa01VQ3HdNON 3
3iVb NITO IVIOI 3
(D
1-1
3
31VU iN3Wlina33d 3
3iva
45
NOT REPRODUCIBLE
-------
Figure 10. Derivation of the carnivore production gain response functions
used in constructing the predator isoclines on the C-H phase plane. Car-
nivore production rate is derived from the curve relating carnivore relative
growth rate to herbivore biomass. If herbivore biomass and, thus, carnivore
relative growth rate were held constant at a series of levels, the relation-
ship between carnivore production rate and carnivore biomass would be linear.
Linear production responses can be generated for each herbivore biomass. The
procedure is illustrated for herbivore biomasses 2 and 3. Derivation of
these problematical carnivore production relationships facilitates construc-
tion of predator isoclines. System-determined carnivore production curves
can be derived from relationships between carnivore and herbivore biomass
defined by the prey isoclines at each light input rate on the C-H phase
plane and from the curve relating carnivore relative growth rate to
herbivore biomass. The relationship between the linear problematical
production responses and the system-determined production curves at three
light input rates (dashed curves) is shown.
46
-------
Ul
o:
i
o
UJ
UJ
cc
o
UJ
-------
Environmental toxicologists seek to understand the effects of toxic
substances on the capacity and performances of natural systems. Toxic
substances introduced into natural systems induce perturbations in the
performances of the systems: observationally, this is why such substances
are first supposed to be toxic. Some possible effects of toxic substances
on the performances of populations composing a defined system and on the
structure of the system can be examined with the sort of deductive argument
presented above.
The effect of a toxicant, T, acting directly only on carnivore C in
the simple system defined in Figure 5C will be examined. To construct the
isoclines on the phase planes representing this system, the effects of
different concentrations of toxicant T on the response functions of the
carnivore are defined. In Figure 11A, the relative growth rate response
of the carnivore is shown to be reduced by increasing concentrations of T
from OT to 21. The carnivore relative growth rate response is used to
construct the carnivore production gain curves (Fig. 10) and, consequently,
toxicant concentration in addition to herbivore biomass parameterizes the
production response functions. In Figure 11B, a toxicant concentration of
2T is shown to be acutely toxic (toxicant induced mortality) to the carnivore.
The remaining carnivore gain and loss response functions (recruitment,
•yield, and nonpredatory losses) are identical to those shown in Figure 9.
In this example, toxicant induced mortality represents an additional loss
response function.
In Figure 12, the phase planes and isocline systems representing the
simple system defined in Figure 5C are shown. In addition to level of
fishing effort,' toxicant concentrations OT, IT, and 2T parameterize predator
isoclines on the C-H phase plane. At a particular level of fishing effort,
increasing toxicant concentration shifts the predator isoclines downward
on the C-H phase plane, as shown in Figure 12. Toxicant concentration, as
well as rate of light input and level of fishing effort, determine the
steady-state biomasses of the populations composing the system and, thus,
the steady-state structure of the defined system. The points locating
the steady-state biomasses of C, H, P, and R at high I, OE, OT (solid
circles), high I, OE, 2T (solid squares) and low I, 90E, 2T (solid triangles)
are indicated on the phase planes in Figure 12.
As yet, only the steady-state performances of a defined system at
different levels of environmental variables have been considered. Environmental
variables in natural systems may never be constant for long enough periods
of time to permit these systems to reach steady-states. The dynamics of
system performances and the response of the system to changes in levels of
environmental factors, including toxicant concentration, are illustrated
on the phase planes shown in Figure 12. Here, trajectories represent the
changes through time in biomasses of populations composing the defined
system.
If the levels of environmental factors I, E, and T are fixed at
high I, OE, OT, from an initial point, 0, simultaneous trajectories of
biomasses on each phase plane (A trajectories) converge upon the points
48
-------
H
u.
1°
B
I
2
I
3
I
4
I
5
OT, IT
Figure 11. (A) Direct effects of toxicant T on the relative growth rate of
carnivore C in relation to herbivore H.biomasses, at concentrations of OT,
IT, and 2T. (B) Direct effects of toxicant on the loss or mortality rate of
carnivore C are shown to occur at 2T but not at OT and IT and are taken to be
linearly proportional to biomass of C.
49
-------
LOW t MED 1 HIGH I
l/l
O
LOW 1 MED t H4GH I
5-
4-
3-
2-
LOW I MED I MICH I
Figure 12. Possible steady-states (indicated by circles, squares, and triangles) and trajectories of
biomasses of carnivore C, herbivore H, and plant P population and light intensity R under different rates
of light input (low, medium, and high) and different level of fishing effort E, when toxicant T is
absent and when it is present at concentrations of IT and 2T.
-------
locating the steady-state biomasses of C, H, P, and R (solid circles).
On each phase plane, the steady-state point is a two-dimensional projection
of a four-dimensional community steady-state point in hyperspace. Similarly,
the A trajectory on each phase plane is also a two-dimensional projection
of a four-dimensional trajectory that converges upon the community steady-
state point.
If, after some period of time, a persistent toxicant (PT) directly
affecting only carnivore C is introduced at a concentration of 2T, then
the trajectories are perturbed and converge upon the points defining the
steady-state biomasses of C, H, P, R at high I, OE, 2T (solid squares).
These are the same steady-state points upon which the trajectories converge
if the toxicant is present initially at a concentration of 2T (B trajectories)
If a nonpersistent toxicant (NPT) directly affecting only carnivore C is
introduced at a concentration of 2T at the point so marked on the A trajec-
tories, these trajectories are perturbed and progress toward the steady-
state points generated by high I, OE, 2T,initially by the paths of the
trajectories resulting from introduction of the persistent toxicant. As the
concentration of the nonpersistent toxicant in the system decreases, however,
the trajectories begin to return to the steady-state points generated by
high I, OE, OT on each phase plane. Through its effects on the carnivore
population, toxicant T indirectly influences the performance of all popula-
tions and the structure of the defined system as a whole.
Now, if the levels of environmental factors I, E and T were fixed at
low I, 90E, 2T, trajectories on each phase plane converge upon the steady-
state biomasses of C, H, P, and R at low I, 90E, 2T (C trajectories).
Under this set of environmental conditions, the prey isocline parameterized
by low I and the predator isocline parameterized by 90E, 2T do not intersect
on the C-H phase plane and the carnivore population is driven to extinction.
Although carnivore C was capable of persisting at a toxicant concentration
of 2T when I and E were fixed at high I, OE (B trajectory on the C-H phase
plane) the carnivore does not have the capacity to persist at 2T when I is
reduced to low I and E increased to 90E (C trajectory on the C-H phase
plane). Perhaps more than anything else, it is the capacity of systems to
persist under toxic perturbation that we should seek to evaluate for
biological systems.
The foregoing example serves to illustrate how the performances of
systems are determined by the levels of environmental factors such as I,
E, and the toxicant T. Changes in the level of any of these environmental
factors alter the performance of the system. Ultimately, in environmental
toxicology, we must attempt to understand how toxicants alter the domain
of system performances. To do so, either theoretically or empirically, we
must evaluate, insofar as possible, the performances of systems under
different sets of environmental conditions. But we must always be aware
that changes in species composition of a defined system, or evolution of
these species, would change the magnitude and form of the population response
functions and thus alter the positions and forms of isocline systems, the
interactive performances of the populations, and the performance of the
system as a whole. Thus increasing the dimensionality of the system to
51
-------
include a greater number of species and kinds of interactions (competition,
commensalism, mutualism) would alter both time-variant and steady-state
performances of the system.
C. THEORETICAL EVALUATION OF THE CAPACITIES OF ORGANISMIC SYSTEMS
Capacity is a theoretical concept that entails all possible performances
of a system in all possible environments. Any organismic system has the
capacity for an indefinite number of kinds and levels of performances, but
the capacity of a system cannot be fully evaluated even theoretically. -
Theoretically and empirically however, partial evaluations of the capacity
of systems to perform--their scopes for performance—can be achieved.
Such a partial evaluation of capacity involves determination of particular
performances of a system over a range of values of particular environmental
factors. Theoretical evaluations of scope for performance should go beyond
empirical evaluations of scope in explicating what the domain of performance
of a system might, be like and how toxicants might alter this domain.
An example of the use of theoretical determinations of scope for
performance to give greater meaning to observational experience is found
in the wprk of Taub (1973). Continuous flow culture of populations of the
alga Chlamydomonas reinhardtii in the presence as well as in the absence
of populations of the protozoan Tetrdhymena vorax permitted Taub to
determine, under different regimes of light intensity, flow rate, and
nutrient concentration, some of the steady state population densities for
the alga and the protozoa. With a simulation model, steady-state density
isopleths for these two populations in relation to light intensity, flow
rate, and nutrient concentration were derived (Fig. 13). Such systems
of isopleths represent steady-state performances of the alga-protozoan
system over a range of levels environmental factors and explicate the
interactive effects of these factors in determining the scope for performance
of the alga, the protozoan, and the alga-protozoan system. Introduction
into this system of a toxicant affecting either or both the alga and the
protozoan would have altered the scope for performance of the system
and the domains of possible steady states under different environmental
regimes.
D. ON DIMENSIONAL, DYNAMIC, AND EMPIRICAL ADEQUACY OF THEORIES
The logical positivistic emphasis of physics, as we earlier noted,
has led to -internal consistency and prediction as the primary criteria
of theoretical adequacy. In mimicking the form of physical theory, biology
has tended to adopt the same criteria. But evaluation of theory mainly on
the basis of these two criteria leaves out too much of what we must expect
of theories, if they are to be adequate for biological explanation and
understanding. Before we go on to elaborate what further external adequacy .
we should expect of biological theory, however, let us consider briefly
what may be meant by prediction, as it is certainly important. First,
prediction as a test of the adequacy of theory is very powerful if, by
prediction, we mean the identification of some previously unknown event,
52
-------
A.
MAX. DENSITY 3.99
B.
60OO
40OO -
2OOO -
6000
MAX. DENSITY 0.57
4000
200O
200
400
60O
20O
4OO
600
MAX. DENSITY 0.09
D.
60OO
MAX. DENSITY 0.689
4000-
2000
200
400 600 0
FLOW
200
400
6OO
Figure 13. Steady-state density isopleths of the alga Chlamydamonas
reinhardtii (A and B) and the protozoan Tetrdhymena vorax (C and D), under
different light and flow regimes when nitrate supply was 0.5 mM (A and C)
and when it was 0.05 mM (B and D) in continuous flow cultures. These
diagrams illustrate population scopes for performance in terms of steady-
state biomasses and environmental variables. After Taub (1973).
53
-------
say a new subatomic particle, that we then find a way of observing or
detecting. To make this sort of prediction possible, it would appear that
a theory somehow logically incorporates important aspects of the underlying
nature of the natural system. But, at least in biology, it is not our
impression that this is what is generally meant by prediction. Rather it
is usually taken to be some statement about the probable value or behavior
of some already known object. If a theory can predict such values or
behaviors of a natural.system under a very wide range of kinds and values
of environmental conditions, prediction again would be a powerful test of
theory, for only a theory having much external adequacy would be able to
do this. But even this is not generally expected by way of prediction as
a use 'Or even a test of biological theory.
What, then, are we to expect in the way of adequacy of theories?
We will consider three aspects of this: dimensional adequacy, dynamic
adequacy and empirical adequacy. Lewontin (1974) employed these terms
in his critical evaluation of the modern genetical theory of evolution,
but we may be extending the meanings of his terms further than he intended.
We do so because it is our deep conviction that theories ought to be
explanations considerably advancing our understanding, and prediction
alone does not insure this. On the other hand, adequate explanation and
understanding tend to insure prediction of both kinds. In considering
these aspects of theoretical adequacy, we must do so in the light of what
we take explanation and understanding to be, and what we take to be the
nature of biological systems. The external adequacy of theories cannot be
evaluated in a vacuum, and the conceptual framework earlier presented can
serve as an aid.
By dimensional adequacy of a biological theory, we mean its
incorporation of sufficient organismic system variables and external or
environmental variables to make possible a reasonably general explanation
of the behavior of natural organismic systems to which the theory is directed.
And we take "sufficient variables" to be naturally important variables,
enough variables, but not too many variables to make general explanation
possible.
In the graphical deductive arguments previously presented, we demonstrated
that addition of environmental variables changed both the behavior and
steady-state structure of a defined organismic system. We could also
have demonstrated that addition of organismic dimensions—or objects—would
change the capacity of the organismic system, as well as its behavior
and structure. Even apart from any deductive argument, we all intuitively
know the importance of addition or deletion of variables from a system
defined to represent some natural system. But natural biological systems,
say biological communities together with their physical environments, may
have thousands of variables. Surely the relatively few dimensions with
which we can deal adequately cannot represent such biological systems.
We believe adequate theoretical representation and explanation of
dimensionally complex natural systems require definition of objects of
these systems at a high enough level to reduce the number of dimensions
54
-------
needing representation. For the biological community, such objects may be
community subsystems, which include trophic subsystems made up of populations.
Only for systems lower than the community level should populations be
represented as objects in a theory. The dimensions that we do include
should make it possible for us to evaluate aspects of the capacity of a
system, not simply its performance under a particular set of environmental
conditions. But no theory of which we can conceive can fully evaluate the
capacity of a natural system.
By dynamic adequacy of a theory, we mean reasonably general represen-
tation of the behavior—both time-variant and steady state—of the natural
system toward which the theory is directed. A theory cannot be dynamically
adequate if it is dimensionally inadequate. But dynamic adequacy involves
the adequacy of relations among variables hypothesized in a theory, as
well as the number and kinds of variables included. For relations, just
as much as dimensions, will determine the behavior, structure, and capacity
of the theoretically defined system, and thus how well this explains the
natural system of interest.
Relational adequacy raises difficulties perhaps even greater than
those raised by the dimensional adequacy criterion. This is because we
may more or less directly be able to perceive the objects we identify as
dimensions, but we can in no ordinary sense observe the relations between,
or the interactions of, the performances of these objects. To suppose
that we can is to ignore all the difficulties of causal explanation
philosophers have been unable to solve. Thus we hypothesize relations
among objects or their performances. The addition of many dimensions to a
theory or simulation model only aggravates the difficulty of hypothesizing
reasonably adequate relations among dimensions.
Now what do we mean by empirical adequacy of a theory? In a narrow
sense, the theorems of a. theory must be translatable into empirical
generalizations, the conceptualized identities of which can be adequately
observed. Otherwise, the theory is irrelevant. But, more generally,
to be empirically adequate a theory must account for the performances we
are able to measure, and do so under ranges of environmental conditions
and system states of general interest. Again, this is only to say that
a theory must to some useful degree permit us to evaluate the capacity
of the organismic system of interest, this ultimately involving empirical
as well as dimensional and dynamic adequacy. An empirically adequate
theory must give meaning to particular performances we measure, and it
must do so by relating these performances to other performances on the
same level of organization as well as on lower and higher levels of organi-
zation. It must do so in a way making it possible for us to know at least
theoretically, the possible applications of empirical results we obtain in
the laboratory or in nature.
55
-------
IV. EMPIRICAL CONSIDERATIONS IN ENVIRONMENTAL TOXICOLOGY
A. ON PROBLEMS OF CREATING BIOLOGICAL SYSTEMS IN THE LABORATORY
Good empirical investigation is important in any science, for empirical
results are our perceptual interface with nature. Good empirical generali-
zations reflect something of nature, even if they fail to order nature as
do universal theories. This makes it important that laboratory research
in biology be conducted so as to yield the most general results possible.
Moreover, such research helps to make clear what there is about nature
that we must develop conceptual frameworks and theories to explain and
understand. We must, however, employ extant conceptual frameworks and
theories, inadequate though they may be, in the design, evaluation, and
application of empirical research. In this part of our consideration of
laboratory research in environmental toxicology, we will deal with problems
of creating biological systems in the laboratory. Then, in following
parts, we will consider determination of organismic system performances
and evaluation of capacities as well as the adequacy of empirical investi-
gations in environmental toxicology.
Before beginning, perhaps we should note again our definition of a
biological system : an organismic system, whatever its level of organi-
zation, together with its level-specific and co-extensive environment,
extending through all states of the organismic system from its origin to
its end (Fig. 2). We can think of such biological systems as occurring at
any level of biological organization, from physiological systems, within
an individual organism, up through biological communities. In this article,
we are concerned mainly with ecological—not physiological—levels of
organization, and we take an individual organism and its environment, a
population and its environment, and a biological community and its environ-
ment to be ecological systems. Ecosystems are generally taken to be
biological communities together with their physico-chemical environments.
There are, of course, intermediate levels of organization with which we
will have to deal.
Before we move on to consider problems of creating laboratory systems .
at different levels of biological organization, let us emphasize general
considerations applying to all levels. Laboratory ecological systems are
physical-biological models. And, hence, as we noted in our earlier discussion
of models, laboratory ecological systems must be evaluated in light of the
objectives for which they are intended. Objectives of biological research
must ultimately relate to natural systems. Thus, we must have some notion
of the natural biological systems we intend to model in the laboratory, or
we would have no way of knowing the meaning of our results or how they
might be applied in reduction of problems of toxic substances in the
environment. . .
The conceptual framework developed by Warren, Allen, and Haefner (MS)
emphasizes the importance of the theoretical concept capacity in the
explanation and understanding of the organismic part of biological systems.
Capacities are specific to particular organismic systems, whatever their
56
-------
levels of organization. Whenever we employ organismic systems of
higher than individual organism levels of organization in our laboratory
studies, especially multispecies systems, we in large part determine the
capacities of the systems by our design. Thus we determine possible
performances. Nevertheless, we must determine or ascertain that the
organismic systems we study have the capacities for the performances in
which we are interested—and that the performances we measure are indeed
those of the organismic system as a whole, and not just some subsystem of
it. Otherwise, we had just as well study the subsystem. Organismic perform-
ances of interest may be persistence, development, structure, replication,
or some other. In environmental toxicology, biological magnification
(accumulation) and degradation of toxicants may be taken to be performances
of organismic systems on different levels of organization.
Now the performances of any organismic system that we measure or
might be able to measure are determined not only by the potential capacity
of that organismic system but also by its environmental system in its
previous as well as present states. In determining the environment of any
organismic system in the laboratory, we further determine what performances
will occur. Thus it is extremely important how we select the environ-
mental factors and their levels to be employed. The behavior and effects
of any toxicant will depend very much on other environmental factors and
their levels. And a range of other environmental factors should be
investigated, so as to at least partially evaluate the capacity of the
organismic system of interest. This is an important way of determining the
meaning and possible applications of performance results.
We can now begin with consideration of difficulties of creating
laboratory biological systems at the individual organism level of organi-
zation, for studies in environmental toxicology. Here, the problems are
mainly associated with creating the environmental part of the biological
system, for an individual organism, as a whole, is in itself a natural
organismic system, owing to reproduction, development, and evolution.
But the difficulties of creating a meaningful environmental system for
individual organisms in the laboratory are real enough. Let us start with
objectives of laboratory toxicological studies of individual organisms.
A common objective is to determine the relative potency of different toxic
substances by employing the individual organism as a "biological reagent,"
in the classical sense of bioassay. Other than standardization of species,
response measurements, and test conditions, there are no great difficulties
in attaining this objective. But rational use of such information is not
a simple matter. Another obj'ective may be to determine accumulation
or decomposition of toxic substances by individual organisms. And,
perhaps most often in environmental toxicology the objective is to determine
the effects of toxic substances on one or more performances of the
individual organism. Because presumably these last sorts of objectives
generally are ultimately related to possible effects of toxicants on species
populations in nature, selection of the environmental factors and levels
as well as the performances to be studied is critical. Not only selection
but also creation of these conditions is difficult. But if these are not
well done, the meaning and possible applications of any results are limited.
57
-------
By individual organism we mean the entire life history, from zygote
through development to reproduction and finally death-. Nearly all individ-
ual organisms, as natural systems, have the capacity for all performances
'upon which this developmental trajectory depends. At this level of organiza-
tion, the problem of creating a biological system in the laboratory is
nearly entirely one of designing and providing an environmental system in
which the life history can be completed. For small organisms having
relatively short life histories, this can very often be done. But obviously,
for many species, facilities and time alone may make this impractical.
Yet, if our studies of the effects of toxicants on individual organisms
are to relate, even indirectly, to the persistence of population in nature,
individual organism performances of reproduction, development, growth
and persistence must be studied.
Environmental Protection Agency scientists have developed and employed
chronic tests in which one or more generations of species of fish or smaller
organisms have been maintained in the laboratory at constant toxicant
concentrations (Mount and Stephan, 1967). In determining toxic effects on
entire life history sequences, such studies are well directed. But under
different sets of environmental conditions, the effects of toxicants on
various life history responses cannot be expected to be the same, and
food availability, temperature, and many other conditions in nature vary
continuously,, So also do concentrations of toxicants and life history
stages exposed. Obviously, all this raises very serious difficulties in
designing individual organism studies so as to make the results applicable
to nature. Somehow we must learn more of the possible ranges of
performances of a kind, under various sets of environmental conditions—
learn more of the effects of toxicants on the capacities of organisms.
Much can be learned about the effects of toxicants on the capacities of
individual organisms for reproduction, development, growth, and persistence
by means of studies not encompassing entire life history trajectories
(Warren, 1971), but such more limited studies do raise questions about
what effects would have occurred if previous or later stages had been
exposed.
A biological population is usually defined as a more or less
isolated group of interbreeding organisms, there being no satisfactory
definition for populations of strictly asexual organisms, of which there
appear to be relatively few. As such, a population should be viewed as an
organismic system on a level of organization higher than that of the
individual organism. Accordingly, biological populations have capacities
and performances that individual organisms do not have. These include -
population growth in numbers or biomass, evolution, and persistence
through many generations. And populations exhibit some structure
or organization: they are not simple aggregations of individual organisms.
The reason for studying populations, rather than some other level of
organization, is interest in population level capacities and performances.
Thus, in environmental toxicology, objectives of population studies could
be to determine effects of toxicants on population growth, production,
evolution, persistence, or the capacities for these performances.
Accumulation or degradation of toxicants by populations as wholes
58
-------
could also be of interest.
We do not, in any ordinary sense, create a biological population in
the laboratory by the mere act of bringing together a group of individuals
of the same species. Rather, if we are to have anything resembling a
biological population, we must provide individuals with environmental
conditions suitable for them to complete their life histories, including
reproduction. And sufficient space and time and suitable environmental
resources and conditions must be provided for growth of a population
through at least several generations. During this time, the population
will develop sex, age, size, and perhaps social structure. And it will
come to have capacities and performances, according to the nature of the
species, the genetic potentials of the original individuals, and main-
tained environmental conditions. For different environmental conditions,
different birth and death rates will come to prevail and laboratory popu-
lations will come to have different structures. And the effects of toxic
substances, as well as toxic substance behavior, will be dependent on
population structure as well as on environmental conditions other than the
toxicant.
Such biological populations are routinely developed in laboratories
for ecological and evolutionary studies, textbooks being replete with
examples. But because of time and space requirements for completion of
many generations and maintaining the populations, most studies have employed
small organisms having short life histories. Among animals, protozoa,
insects, and small Crustacea have been most studied in this way, although
there have been studies on small rodents and fish as well as other groups.
Studies of the capacities and performances of populations of small organisms
in the laboratory could be of much more importance than they have been in
environmental toxicology. There is much precedent for such studies, and
a sound empirical basis, but there is also much to warn us of the difficulties
inherent in their conduct and interpretation.
Here again, as with the individual organism, the difficulties center
mainly around our being able to provide suitable environmental conditions
for the completion of life histories. Even when this is done, and popula-
tions develop and persist for many generations, possible meanings and
application of results depend nearly entirely on how well the natural
environment of any studied population has been modeled. We can hardly
replicate well all physical and biological conditions of a natural
population's environment. We may come most near to doing so with some
phytoplankton and zooplankton species and certain grain weevils. But for
larger species having more complex environments, introduction of environ-
mental factors such as several prey species, competitors, and predators in
a realistic way presents serious difficulties. Nevertheless, there have
been modest successes in some work of this sort, and models are always
justifiably simplifications of nature. We would only emphasize one
further matter: the realized capacities that develop and the performances
that occur in a laboratory population are specific to the prevailing
laboratory environments. We should explore population capacities and
performances over as wide a range of selected environmental conditions as
59
-------
possible, if it is our intent, as it usually must be, to apply our results
to nature. And in such exploration we must, insofar as'possible, account
for both time-variant and steady-state behavior of population level
performances.
/
Possible creation in the laboratory of mu.ltispeci.es eoologiaal systems
that have level-specific capacities and performances, making them systems
in their own right, presents very great difficulties. Even the results of
classical predator-prey and competition studies, in which at most a few
interacting species populations have been maintained in the laboratory,
have very doubtful meaning and have not been related well to natural systems.
At multispecies biological system levels, we not only bring forward the
difficulties associated with providing suitable environments for completion
of life histories and maintenance of populations—which we have at
individual and population levels of organization—we have actually to
compose a meaningful organismic system of two or more interacting species
populations. Even if we are successful in creating a multispecies system,
we in large part, by the species we bring together, determine what the
capacities and thus the performances of that system will be.
We defined a biological community, in the organismic sense, as the
sere composed of all successional or developmental stages of the community^
We suppose such high-level organismic systems to exist in nature, or we
would have little reason to talk of them. On the basis of this supposition,
biological communities must have level-specific capacities and performances,
however well we are able to observe them. The major performances of a
biological community can be taken to be development (or succession) ,
structure., and persistence. The effects of toxic substances on
any of these can have serious implications for man. And movement,
accumulation, and degradation of toxic substances by biological commu-
nities is also a matter of great concern, one that has led to many studies
employing laboratory models.
Our reasons for studying a particular level of biological organi-
zation presumably are based on interest in capacities and performances of
systems having that level of organization. Very few if any laboratory
systems can be believed to represent ecosystems, for this implies that
biological communities, with all their complexities, capacities, and
performances have been created. Many attempts, with some success, have
been made to create multispecies systems for toxicological studies in the .
laboratory, most of which have involved relatively few species and certainly
can represent, at most, very low levels in community organization.
One objective of multispecies systems studies in environmental toxi-
cology is assay of the relative tosciaity of substances. There is, however,
considerable difficulty in experimentally replicating multispecies systems
in a way that would be necessary for them to serve as "reagent" organismic
systems for assaying the relative toxicities of substances.
The major objective of most of the published work employing multispecies
systems in environmental toxicology has been determination of transport^
60
-------
distribution, and degradation of toxic substances. Metcalf, Sangha,
and Kapoor (1971, p. 709, p. 711) state: "We need a realistic laboratory
method for screening proposed new pesticides for their environmental fate."
And, "The ultimate purpose of the model ecosystem is to be used as a single
living unit for in-depth studies of environmental biodegradability."
Isensee -and Jones (1975, p. 668), with respect to their laboratory ecosystem,
note: "This system is not designed to determine the effects of a chemical
on the organisms (though some primary effects can be assessed) but rather,
how does the chemical behave when subjected to likely environmental
conditions." And Gillett and Gile (1975, p. 1) state that "The need to
know or project accurately potential effects of chemicals deliberately or
accidentally added to the environment forms the basis for the development
of laboratory ecosystem simulators."
In our own laboratory we have used simple multispecies systems to
evaluate the effects of such factors as temperature, dissolved oxygen,
ammonia, pulp and paper industry effluents, and other toxic materials
on the behavior, food utiliztion, growth, development, and survival of
individual fish. The use of multispecies systems for this purpose permits
evaluation of the performance and capacity of individual organisms under
conditions perhaps more like those in nature than are usually provided in simple
aquarium studies. In many cases, but not always, these studies have been a
part of broader investigations directed toward effects on populations and on
the multispecies systems as wholes.
Among others, Taub (1973) has conducted multispecies system studies
having the objective of evaluating direct and indirect effects of toxicants
on populations of algae and Ltxphnia. Some of our own studies on fish
population in laboratory stream systems have had the objective of deter-
mining direct and indirect effects of various toxic and other environ-
mental changes not only on the production of populations but also on
the capacities of biological systems to produce them. These studies
have in some cases been directed toward determination of the effects of
environmental changes on the persistence, development, and structure of
laboratory communities as wholes (Warren and Davis, 1971). Such
community-level objectives have been implicit in studies of other
investigators.
The laboratory systems developed by Metcalf (1974) and his colleagues
(Metcalf, Sangha, and Kapoor, 1971, and many papers following this one)
are designed primarily to evaluate accumulation and degradation of
toxic substances. These systems are contained within glass aquaria
measuring 10 x 12 x 20 inches and illuminated with 5000 ft candles of
light 12 hours each day. Sand is formed into a sloping surface and a
standard water added so as to represent both terrestrial and aquatic
habitats. After 1 day, Sorghum halpense seeds are planted, and 10
Physa snails, 30 Dxphnia magnat and a few strands of the alga
Oedogonium cardiaaum are introduced. In 20 days, when the Sorghum
plants have reached about 4 inches in height and are treated with, radio-
labeled pesticide, 10 larvae of the salt marsh caterpillar, $st^gmene
aarea are added. On the 26th day, 300 larvae of the mosquito Culex
61
-------
pipens quinquefasciatus, and on the 30th day 3 mosquito fish, Garribusia
affinis are introduced to complete the laboratory model. The mosquito
fish readily consume all remaining Dzphnia. and mosquito larvae, and the
experiment is terminated at the end of 33 days. Samples of sand and water
are collected periodically during the experiment and biological materials
are collected at the end of the experiment for analysis of concentrations
of the introduced toxic compound and its metabolites.
Isensee and his colleagues (Isensee, Kearney, Woolsen, Jones, and
Williams, 1973) have designed and employed laboratory models essentially
like those of Metcalf except that no terrestrial phase has been provided.
Aquaria holding 4 liters of water were kept in a greenhouse at about 22 C.
Dzphnidf OedogoniiOTij and Fhysa were first introduced. After 29 days,
by which time the Etiphnia were found to have increased from about 30 to
100, 2 mosquito fish were added to each of the aquaria. The experiment
was terminated 3 days later, after the fish had consumed all the Dzphnia.
Toxic substances were introduced either directly into the water or first
adsorbed onto soil that was added. Analyses for concentrations of these
substances in water and organisms were made during and at the completion
of the experiments.
Gillette and Gile (1975) have developed quite elaborate physical
systems for containing terrestrial organisms and controlling environmental
conditions. Air flow and humidity can be controlled and water is provided
by a sprinkler system and a "spring" system. Light intensity is controlled
even to the extent of providing short "dawn" and "dusk" periods, and
there is careful monitoring of temperature at various points within the
chambers, which are in a temperature controlled room. The chambers them-
selves are 1 meter long and 0.75 meter wide. A considerable variety of
plant and animal species is maintained in these terrestrial laboratory
models, including the vole Microaaudus oanioaudus and the quail
Excdlfaatoria chinensis, food for which had to be supplemented. By the
end of an experiment, 30 to 45 days, "The pregnant female vole literally
destroys the terrarium." Cole, Sandborn, and Metcalf (1976),who included
the vole Mic^otus ochvogastu in a terrestrial model with corn plants and
other organisms during the last 5 days of an experiment, noted that "it is
not uncommon to find that all other organisms of the system have been
consumed" by the vole.
Now, before we go on to review the design of laboratory models employed
in studies having other objectives, let us only in very general terms
evaluate the designs of multispe'cies systems directed toward determination
of the transport, accumulation, and biodegradation of toxic substances.
The sorts of laboratory models we have just characterized do not satisfy
the rather severe criteria for organismic and environmental systems that
we have argued should be met if we are indeed to have multispecies labora-
tory systems. The species of organisms included in these laboratory
models are not adapted to one another so as to lead to the development of
multispecies systems capable of persisting under the environmental condi-
tions provided. The scientists employing these laboratory models are
generally aware of this. This awareness is implicit in the experimental
62
-------
protocols in which fish or voles are introduced into the models only
during the last few days of experiments, the models not having the capacity
to support these vertebrates for much longer. The awareness is repeatedly
made explicit in comments to the effect that the fish or the voles consume
all their food organisms in a few days. Even in the larger and more
productive biological model developed by Gillett and Gile (1975), supple-
mental feeding was necessary to maintain the quail, and in 30 to 45 days
single voles were noted to have destroyed the plants and animals in their
environments. These laboratory models, at least so long as the verte-
brates are included, are not persistent systems and may not reflect
reliably the behavior and effects of toxic substances in natural systems.
One of the major reasons systems having some capacity for persistence
under provided environmental conditions have not been created is failure
to insure that, at each step or link in a food chain or web, sufficient
energy and material resources will be available. For the plant trophic
level, sufficient light energy and nutrients must be provided not only for
plant maintenance but also for plant growth adequate to provide energy and
materials for succeeding trophic steps. And at each of these succeeding
steps, the animals selected for inclusion must be capable of producing
enough material under prevailing conditions to support those following.
Ideally, both reproduction and growth capacities and performances of
included organisms should be sufficient to support other organisms depend-
ing upon them; but multispecies systems can sometimes be designed so as to
persist for useful periods of time mainly on the basis of the growth and
production of their organisms. But neither growth nor reproduction of
food chain organisms will be adequate to support larger animals such as
fish, birds, and voles in laboratory models having such small physical
dimensions as we have considered, available light energy and plant produc-
tion being too low for this. Rather large and carefully designed laboratory
ecosystems are necessary if large animals are to be sustained.
We wonder if the kind of information on behavior and effects of
toxic substances obtained from short-term studies of nonpersistent systems
cannot be obtained more simply and comprehensively from studies of individual
organisms or groups of individual organisms of the same species. Individual
organisms can more simply be held in toxicant contaminated medium and fed
contaminated food organisms than in even partially self-sustaining systems.
Survival, growth, and even reproduction rates as well as accumulation and
degradation can be determined over a wider range of toxicant concentrations
and energy and material utilization rates. Thus lethal and sublethal toxicant
effects and accumulation and degradation can be evaluated and interrelated.
Good information of this sort can be incorporated into theoretical deductive
systems and models and so be extended to prediction of multispecies system
effects, perhaps more reliably than prediction can be made from all but the
most sophisticated laboratory ecosystems.
Let us now move on to consider laboratory model studies having some-
what different objectives. The work of Taub (1973) and her colleagues
on algae and protozoa, which we considered earlier, is exemplary in comple-
menting well-designed studies of multispecies systems with generalization
f
63
-------
by means of a reasonably adequate simulation model (Fig. 13). Work of
this sort amounts to a rather extensive exploration o£ the(theoretical
capacity of a biological system. Introduction into this system of a
toxicant affecting either or both the alga and the protozoan would have
altered the scope for performance of the system as a whole and the domains
of steady states under different environmental regimes.
Many of our own studies employing laboratory systems have had the
objective of determining the effects of toxic substances and environmental
factors on the capacity of these systems to produce products of interest,
especially salmonid fishes. We began laboratory stream community studies
about 20 years ago and were soon able to maintain groups of salmonids
feeding on insects and other invertebrates produced in these streams for
periods as long as we might wish, although only a few experiments were
allowed to continue for more than about 6 months. But even though these
were moderately persistent systems having capacities and exhibiting per-
formances in which we were interested, they did not .entirely satisfy our
criteria for multispecies laboratory ecosystems. In particular, comple-
tion of trout life histories requires two or more years, and anadromous
salmon must go to the sea and return before reproducing. Because of the
importance of reproduction in determining the persistence, production, and
yield of fish populations exposed to toxicants, we began studies with
guppies, a fish species capable of reproducing fairly rapidly in the
laboratory.
Some studies were directed toward evaluation of the effects of dieldrin
on the age-specific fecundity and survival of individual guppies, such
information making it possible to theoretically determine effects on the
capacity of guppy populations to increase in number (Roelofs, 1971).
Laboratory population studies with guppies were conducted by Liss (1974).
He maintained populations of guppies for 15 consecutive months and exploited
them at different rates up to 60 percent per month to determine the effects
of dieldrin on population biomass, number, and yield to exploitation.
Biologists at our laboratory are now evaluating the effects of dieldrin on
the structure and persistence of multispecies systems composed of populations
of guppies, amphipods, copepods, and other invertebrate species. The guppy
populations in these systems are being subjected to different rates of
exploitation. To insure persistence of the systems prior to introduction of
toxicant, invertebrate, prey species that could support guppy populations had •
to be found and an adequate food supply and protective habitat for the
invertebrates had to be developed. These systems have been in operation for
one year.
In the 1950's we began studies on the effects of pulp and paper
mill effluents on organisms in laboratory streams. It soon became
apparent to us that we did not understand laboratory stream communities
well enough to be employing them in the evaluation of toxic materials.
Thus, we discontinued such evaluation and undertook several years of study
on the dynamics and productive capacity of laboratory stream communities
including bacteria, algae, insects, and other invertebrates, and different
species of fish (see Warren and Davis, 1971, for a review).
64
-------
Only after we had achieved a reasonably good empirical understanding
of these systems did we return to the problem of evaluating the effects of
pulp and paper industry effluents on the structure, dynamics, and fish
productive capacity of laboratory stream communities. Results of this
work made it possible for us to estimate very closely those concentrations
of pulp and paper industry effluents which did not decrease fish production
in large outdoor stream channels (Warren et al0, 1974). Our own experience,
then, has been that it is important to achieve good understanding of
laboratory systems before employing them in evaluation of complex environ-
mental problems,
• On the basis of this understanding, we have evaluated the effects of
temperature, dissolved oxygen, chloramines, pesticides, and heavy metals
as well as pulp and paper industry effluents on community structure and
the capacity of these systems to produce fish (Warren, 1971; Warren and
Davis, 1971; Bisson and Davis, 1976; Seim et al. 1977). But generali-
zation of the findings of such empirical investigations so as to
apply them with more confidence to the natural systems where the
problems occur requires adequate conceptual structures. Limitations of
the applicability of our empirical investigations convinced us of the need
for conceptual frameworks and theory.
B. EMPIRICAL DETERMINATION OF ORGANISMIC SYSTEM PERFORMANCES
All we can ever observe or measure, however directly, are particular
performances of organismic systems, whatever may be their levels of organi-
zation. All else we attribute to systems or their performances by means of
conceptual structures of various kinds and degrees of abstractness, general-
ity, and invariance (Fig. 1). We cannot observe the capacities of natural
systems, nor their operation, nor their functions or purposes. On this,
important philosphers agree. But they also agree on the importance and
pragmatic rationality of so conceiving of systems that we can explain them
in these ways, thus making understanding possible. Perceptual experiences—
the system performances we observe—are in continuous flux. Such
experience, in itself, has no meaning—is not understandable. We try to
find general and invariant ways—universal ways—of making perceptual
experience understandable. Environmental toxicology cannot escape this
imperative. Otherwise, it cannot give meaning to its empirical results and
thus cannot reliably apply these to solutions of potentially very serious
problems of toxic substances in man's environment.
In this section, we intend to say something of the performances we
observe, the levels of system organization with which they appear to be
associated, and the problem of giving meaning to or understanding observed
performances so as to know something of their generality, invariance, and
possible application to problem solution in environmental toxicology. In
the next section, we will consider the empirical evaluation of capacity, a
very partial and yet effective approach to determination of the domain of
applicability of performance measurements.
65
-------
First, it is important that we know, insofar as possible, the level
of organization having the performance we are interested in and intend to
measure. For laboratory research, failure to associate performances with
the organizational level from which they derive may lead to attempts to
create far move complex systems than necessary to obtain desired per-
formances, to incorrect interpretation and application of results, and to
loss of opportunity to more adequately evaluate the capacity of the sub-
system actually having the performances of interest. We have considererd
organismic systems at the individual, the population, the multispecies
system, and the community levels of organization, though we have concluded
that little can adequately be done in the laboratory at a level of the
biological community. For the individual organism, development, growth,
reproduction, and persistence are major performances that can be deter-
mined. Population performances are growth (change in number or biomass),
structure, evolution, and persistence. And multispecies systems may
exhibit some sort of development, as well as structure and persistence.
But there are real difficulties in interpreting single or even series of
performances of systems on any of these levels of organization. Let us
expand on this topic only for one sort of performance, say growth, at the
individual level of organization: the same problems of interpretation
exist for any performance of any organismic system on any level of
organization.
Now were we to have an individual organism in the laboratory, under a
given set of environmental conditions and fed a given amount of food, we
could measure its growth. The particular growth performance measured
could be "explained" as a result of the food eaten under the prevailing
set of environmental -conditions. But what more could we say? We know
that, in all probability, had the amount of food or the quality of food
fed been different, the growth performance measured would have been dif-
ferent. Moreover, had any important factor in the set of environmental
conditions been at a different level, the growth performance would prob-
ably have been different, even for the given amount and quality of food.
And had the animal been at a different size or stage in its life history,
all else being the same, the measured growth performance would have been
different. Or had the animal been exposed to some different set of
environmental conditions before the growth experiment began, still all
else being the same, it is quite likely a different growth performance
would then have been measured. There is virtually nothing of a general
and invariant sort that we can say about a single measurement of perform-
ance. We do, not know how to apply such a result to nature, where we "know
.environmental conditions, are different and in continuous flux.-
This simple and obvious example is perhaps trivial. And yet it
pertains to all single measurements of performance of any organismic
system: whatsoever. Moreover, the problem is compounded when we attempt to
interpret single measurements of performance of organismic systems on
levels of organization higher than the individual. We want and indeed
need to be able to apply our laboratory results to the solution of prob-
lems in environmental toxicology. But this cannot reliably be done with
single measurements of performances under given conditions, no matter the
66
-------
complexity of the systems we study8 Indeed, the very complexity of the
systems we study may delude us into believing our measurements of
performance are more significant than we have any reason to believe.
Perhaps we should further note that, were we to be studying the effects,
accumulation, or degradation of a toxicant in an experiment something like
the one described, a single measurement of any of these under any of the
fixed conditions would have no more meaning than a. single measurement of
growth performance. -For changes in rate of food consumption and in environ-
mental conditions would change the metabolic state of the organism, the
opportunity for accumulation of the toxicant, and probably effects, accumu-
lation, and degradation of the toxicant. And all such changes would carry
through to affect related performances of higher level systems, were the
individual organism to be a part of these.
The results of an experiment conducted by Morgan (1972) on the
effects of Arochlor 1242 on population growth of a species of algae are
presented in Figure 14. We introduce this here to illustrate, for a
population level performance, the importance of successive measurements
over a sufficient period of time when time-variant performances are being
evaluated. Had measurement of the density of each population been only at
day 3, day 10, day 15, or day 22, one would reach different conclusions on
the effects of this toxicant on population growth, under the experimental
conditions. But because Morgan, through a series of measurements, deter-
mined the time trajectories of population growth, much more complete inter-
pretation is possible. The effects of the toxicant on population density
was time variant until equilibrium conditions were reached. Had this
been an open system providing continuous renewal of nutrients at different
concentrations and p'erhaps removal of algal cells, or had light or tempera-
ture conditions been different, time-variant and equilibrium densities
and observed effects of the toxicant would in all probability have been
different. This experiment is a partial evaluation of population capacity
for growth in the presence of a toxicant. Even so, population performances
would need to be determined under a wider range of conditions, if the
results were intended to relate reliably to possible effects of the
toxicant in nature.
The behavior and effects of chlorinated hydrocarbon pesticides
introduced in natural and cultural ecosystems have provided much empirical
basis for concern abut the possible persistence and accumulation or
degradation of toxic substances employed by man. The emphasis of much of
the work directed toward evaluation of transport, accumulation, and degra-
dation of such substances has been on estimation of ecological magnification
and biological degradability indices. The ecological magnification index
is generally defined as the ratio of the concentration of a substance in
any biological element of a system to that present in water or soil,
although partial indices may be defined as the ratio of concentrations
between any two elements in a system transport pathway. The biodegradability
index is generally defined as the ratio of the concentration of polar
metabolites to that of the remaining parent compound, this ratio being
determinable for an individual organism, a population of organisms, or a
67
-------
r
^
\
>J
2
3
CONTROL
O— SOLVENT CONTROL
Jr- 0.2p.p.m.
- 2.Op.p.m.
- 2Op.p.m.
8
12
DAYS
16
2O
Figure 14. Effects of the addition of 0.2, 2, and 20 ppm Aroclor 1242 on
the growth of Chlamydomonas reinhardtii. Each point represents the mean of
two observations whose individual values are indicated as the range. After
Morgan (1972).
68
-------
laboratory ecosystem as a whole. Accordingly, the values obtained for
either of these indices will be dependent on just what level of system
organization is included in their definition. Gillett and Gile (1975)
suggest that persistent—and implicitly destructive— toxicants have,
on the basis of field studies, ecological magnification indices usually
greater than 1,000 and biodegradability indices usually less than 0.5;
whereas presumably less harmful toxicants have ecological magnification
indices usually less than 100 and biodegradability indices usually greater
than 1. Much of our concern must, then, be with the reliability and
generality of estimates of ecological magnification and biological degra-
dation indices obtained from laboratory systems, and with how reliable these
indices may be in identifying actually or potentially harmful compounds.
As we have stated earlier, by design we determine the capacities and
thus the possible performances of multispecies laboratory systems. Levels
of all performances, including accumulation and degradation, will be
dependent upon the species included in the system and their interactions,
the state of the system including levels and rates of utilization of energy
and material resources, the metabolic states of the organisms, and the levels
of environmental factors, including toxicant concentration, that are being
tested. In Table 1, some examples of the differences in ecological magni-
fication indices obtained from systems with different designs and experi-
mental protocols are presented.
Others have noted various aspects of the difficulties to which we allude.
Matsumura and Benezet (1973) state that: "It is apparent from these data
that the reaction of biological concentration is greatly influenced by the
external conditions and the design of the experiment, the physical and bio-
logical nature of the organisms, and the chemical characteristics of the
pesticide." Isensee et al. (1973) note that: "Use of nonrepresentive rates
[of toxicant introduction] may result in erroneously high or low bioaccu-
mulation ratios which would be of no value in extrapolating laboratory
results to actual environmental conditions." And Isensee and Jones (1975)
state that: "The bioaccumulation ratios obtained in this study were higher
than the concentration factors obtained by Matsumura and Benezet (1973).
However, the experimental conditions used in the two studies were considerably
different—i.e., we used a longer exposure time, different organisms, a
larger system design, and a wider range of water concentrations."
This raises the problem of what designs3 protocolss and conditions
are sufficiently general to provide reliable estimates upon which to
base important decisions pertaining to the use of toxicants. Furthermore,
for a system of a particular design, we believe it is important to evalu-
ate the performances of the system, including transport and accumulation,
over a range of environmental conditions—to explore the domains of
system and subsystem performance—if we are to apply our results with any
confidence to nature. We will discuss this in more detail in the follow-
ing section.
Finally, transport, fate, and degradation of toxic substances are
important in explanation and understanding of effects of toxicants on
69
-------
natural systems, but ultimately it is effects that must concern us,
Single estimates of ecological magnification or biodegradation cannot
possibly account for the numerous kinds and magnitudes of effects of toxic
substances on complex biological systems.
70
-------
CHEMICAL
ECOLOGICAL MAGNIFICATION
REFERENCES
DDT
Mosquito
16,765:21,571
8,182
Fish
218;306
84,500
Hatsumura and Benezet, 1973 (Aquatic systems')
Metcalf, 1974b(Terrestrial-aquatic systems)
Dieldrin
Snail
114,935
61,657
Fish
6,145
2,700
Sanbom and Yu, 1973 (Terrestrial-aquatic systems)
Metcalf, Kapoor, Lu, Schuth, and Sherman, 1973
(Terrestrial-aquatic systems)
Hexachloro-
benzene
1,129
201
Mosquito
2,622
144
Snail
2.672
1,247
Fish
1,166
287
Lu and Metcalf, 1975 (Aquatic systems)
Metcalf, Kapoor, Lu, Schuth, and Sherman
1973 (Terrestrial-aquatic systems)
Dioxin
Danhnia
49;2198
7,800-48,000
Fish
54
1,000-63,300
Matsumura and Benezet, 1973 (Aquatic systems)
Isensee and Jones, 1975 (Aquatic systems)
Chlordene
Heptachlor
Heptachlor
epoxide
Snail
4,167
53,038
1,841
37,153
2,781
66,462
Fish
465
1,122
1,143
3,820
1,324
4,888
Lu, Metcalf, Hirwe, and Williams, 1975
Aquatic systems (3 day test)
Terrestrial-aquatic systems (33 day test)
Aquatic systems (3 day test)
Terrestrial-aquatic systems (33 day test)
Aquatic systems (3 day test)
Terrestrial-Aquatic systems (33 day test)
Table 1. Ecological magnification indices (concentration of chemical in organism/concentration of
chemical in water) obtained from laboratory multispecies systems having different designs and
experimental protocols.
-------
C. EMPIRICAL EVALUATION OF ORGANISMIC SYSTEM CAPACITIES
We have been leading, from conceptual framework considerations through
theoretical considerations and now through empirical considerations, to
the fundamental importance of capacity to organismic systems and our
explanation and understanding of them. In presenting our conceptual
framework, we endeavored to make clear our abstract concepts of potential
capacity and realized capacity of organismic systems for performance.
Then, with isocline theory, we attempted to theoretically show how
organismic system characteristics determine the domain of environmental
dimensions and values within which an organismic system can perform and
persist, and how this domain might be influenced by introduction of a
toxicant. And we have just completed consideration of the very limited
meaning that can be given single empirical determinations of performances,
even steady-state ones.
Now we must consider how and to what extent we can hope to empirically
evaluate the capacities of organismic systems for performance. The very
nature of our notion of potential capacity—all possible performances in
all possible environments—and our notion of realized capacity—all
possible performances at any state of the organismic system—make clear
that neither theoretically nor empirically can we fully and directly
evaluate either the capacity of a natural system or the capacity of a
laboratory system (Fig0 15). But the fundamental importance of capacity
to organismic systems and our explanation and understanding of them demand
evaluation of capacity to the extent possible.
We will begin by making a clear distinction between our abstract
concept of capacity and a theoretically or empirically determined scope
for performance. We will take scope for performance to be a series of
performance values determined over a series of values of selected environ-
mental variables. Thus scope for performance^ so defined and determined,
is an indirect and very partial evaluation of capacity of an organismic
system. In principle, such scopes for performance are determinable for
organismic systems from physiological to community levels of biological
organization. In practice, this can usually be done most adequately for
individual organisms. We will begin with consideration of empirical
determinations of scope for performance of individual organisms.
Were we to think of the total rate of energy metabolism by an
individual organism to be a performance of that organism as a whole, then
we could define metabolic scope for performance. F. E. J. Fry (1947)
defined (metabolic) scope for activity as the difference between the
active and resting metabolic rates of an animal under given environmental
conditions. In scope for metabolic performance, we extend his meaning to
include the series of differences between active and resting metabolic
rates over a series of values of selected environmental variables. Fry's
way of classifying environmental factors into controlling factors (i.e.
temperature), limiting factors (i.e. oxygen), and stressing factors (i.e.
toxicants) could prove to be a very powerful way of determining principles
of environmental effects on organismic system capacities, without our
72
-------
CAPACITY OF NATURAL SYSTEMS
All possible performances
in all possible environments
Utilization of Conceptual
Constructs in Application
of Laboratory Results
to Nature
CAPACITY OF LABORATORY SYSTEMS
All possible performances in
all possible environments
More Extensive
Evaluations of the
Capacity of Systems
•Scopes for Performance
Determination of a
Particular Performance
Under a Particular Set
of Environmental
Conditions
Figure 15. Diagrammatic representation of theoretical and empirical partial
evaluation of capacities of laboratory ecological systems by determination
of scopes for performance as sets of performances under different environ-
mental conditions.
73
-------
having to study all possible environmental factors0 Determination of
scope for performance has already been shown to be a very useful partial
evaluation of organismic system capacities. Figure 16 illustrates how
metabolic scope for performance is primarily determined by a controlling
factor, temperature (Figc 16A). A limiting factor, oxygen, would reduce
scope and shift the maximum (Fig. 16C). And a stressing factor, for
example some toxicant increasing resting metabolic rate, could further
reduce metabolic scope for performance (Fig. 16D). Now any one of the
shaded areas shown in Figure 16 is a scope, for performance over the ranges
of environmental factors for which it was determined. All of these
scopes, along with all performance levels involved, are implicit in and a
part of the realized capacity of any organism. Under other environmental
conditions, other performances and scopes could have been determined, all a
part of realized capacity. And if the organism had a different previous
environmental history, its realized capacity and determined scopes for
metabolic performances would in all probability have been somewhat dif-
ferent: hence the need for the concept potential capacity (Fig. 3).
Warren and Davis (1967} and Warren (1971) extended Fry's (1947)
meaning of scope for activity to scope for growth performance* this
requiring the development and use of bioenergetic concepts we need not go
into here. For even without doing so, we think the different shaded areas
in Figure 17 illustrate sufficiently how changing temperature and food
availability determine the scope for growth of an animal. In the context
of environmental toxicology, there are a few pertinent matters we should
emphasize. Were a toxicant to be added as an additional environmental
factor, according to its level and mode of action, it could alter metabolic
state, food consumption, or food utilization for growth. And such effects
could be expected to be different at different levels of food availability,
temperature, or other environmental factors. We, at least, would also
expect accumulation and degradation of the toxicant to be altered not only
by its concentration in food and medium but also by other factors. In the
time-variant and steady-state behavior of multispecies systems, changing
actions and effects at the individual level must be involved. But it is
not clear to us that this is always taken into consideration.
For the level of the individual organism, there exists an extensive
literature on the interaction of environmental variables, including toxicants,
on scopes for survival, development, growth, and reproduction. We would
only note that information on effects of interacting environmental variables
on entire life history trajectories of particular species has received
very little attention, with some noteworthy exception. Much of the best
of this literature pertains to insects (Andrewartha and Birch, 1954).
Because of the difficulty of empirically evaluating the effects and behavior
of toxicants in populations and multispecies systems we must find better
ways of relating individual level performances to their possible effects
on the capacities and performances of higher level systems (Fig. 4).
Most of the literature on scope for survival of individual organisms,
as series of survival performances under different environmental conditions,
has perhaps understamdably been under acute rather than chronic conditions,
especially when toxicants have been evaluated (Fig. 18)„ For understanding
74
-------
£
I
a
i
Active Metabolic Rate
(A)
Standard
Metabolic Rate
Lethal factors reduce range
of activity but dp not
affect metabolic
scope
Incipient
Lethal Levels
Limiting factors reduce the
scope, limit upper range, ^ —
and usually displace, '
maximum /
Cost of physiological regulation
mi'coc minimum matn- ~~ ~" ~~
raises minimum meta
bolic rate and re-
duces both range /
and scope
Temperature
Temperature
Figure 16. The influences of a controlling factor such as temperature (k),
lethal levels of the controlling factor (B), a limiting factor such as
dissolved oxygen CC), and the cost of additional physiological regulation as
might be. caused by a toxicant (D) on the active and standard metabolic rates
and on|the metabolic scope for performance of a hypothetical poikilothermic
animal'such as a fish. After Fry (1947).
75
-------
X
o
o
u
-X
X
o
o
0)
•»—
D
E
O
0>
,
o>
• o>
FOOD AVAILABILITY MODERATE
FOOD AVAILABILITY LOW
Temperature
Temperature
Figure 17. Theoretical effects of temperature change on the food consumption,
energy budget, and scope for growth performance of a hypothetical poikilo-
therraic animal having food available in different amounts. After Warren .
(1971).
76
-------
10,000
5000
0)
c
2000
- 1000
o
C
o
500
200
100
Total hardness of
water as
O I2mg/l
A 50
D 320
0.5
2 5 10 20
Zinc concentration (mg/l)
50
100
Figure 18. Relationships between median survival time of rainbow trout
and concentration of zinc ion in water having different total hardness.
After Lloyd (1960).
77
-------
problems of this sort in nature, such work is very important, but
the effects of toxicants and their interactions with other environmental
factors under acute conditions cannot generally be extrapolated to include
chronic conditions, which probably present the most general problem. Many
studies that have provided valuable information on the effects of inter-
acting environmental variables on the scope for survival of organisms at
different life history stages, such as shown in Figure 19 for crab larvae,
ought not to be ignored in empirical and theoretical consideration of
problems associated with toxic substances. For, even in the absence of
toxic substances introduced by man, animals are perhaps quite generally,
at one or another life stage, living under conditions of limitation or
stress, as we believe developing salmonid embryos to be (Fig. 20). Under
such conditions of stress* we should expect relatively low levels of
toxicants to decrease scope for performance much more than we observe
under the generally otherwise favorable conditions of toxicity experi-
ments. Toxicants can certainly decrease the scope for growth of animals,
but factors such as temperature (Fig. 2 1) and oxygen concentration (Fig.
22), so important in determining growth in the aquatic environment, are
not usually introduced as additional variables in toxicity experiments on
growth. Moreover, the effects of toxicants as well as temperature and
oxygen on growth are very generally a function of food availability, which
is rarely considered.
The effects of toxicants on survival, development, growth, and
reproduction of individual organisms must generally affect the scopes for
performance of their populations, although the relationships involved are
extremely complex. Population growth in numbers, for example, is determined
not only by individual fecundity and survival in different life history
stages but also by the age structure of a population, which is a complex
function of the two individual kinds of performance. With constant environ-
mental conditions and exponential population growth, populations theore-
tically come to stable, or steady-rstate, age distributions characteristic
of particular sets of environmental conditions. This theoretical proposi-
tion has been widely employed in evaluation of population scope for increase
under different sets of environmental conditions (Fig. 23). Unfortunately,
we can rarely if ever expect the assumptions on which this proposition is
based to prevail in nature. Even so, it provides one of the few quanti-
tative ways we have for thinking about relationships between individual and
population level performances. So long as the assumptions are kept in mind,
toxicity studies could well take advantage of this approach to evaluation
of population capacity for increase, much as Roelofs (1971) did for effects
of dieldrin on the guppy.
The yield of exploited fish populations to man is another population
level performance all too generally affected by man's indiscriminate use and
wasting of toxic materials. Silliman (1968), with experimental guppy popu-
lations, showed how equilibrium yield curves are dependent upon food resource
availability for the exploited population (Fig. 24). Equilibrium yield
curves as well as recruitment and production curves have been shown to be
determined by energy pathways, competition, temperature, toxic substances,
and other dimensions of the productivity of the system (Warren, 1971; Liss
and Warren, MS).
78
-------
r
r
I
20\
I
10 12.5
20 J 26.7
SALINITY (ppt)
J/. / 55
Figure 19. Percentage isopleths representing scope for survival of the
fourth, zoeai stage of larvae of the crab Sesarma oinereum in relation to
temperature and salinity of water. After Maguire (1973).
79
-------
3 4 5 6 7 8 9IOII
Dissolved oxygen (mg/l)
Figure 20. A three-dimensional diagram showing the influence of both oxygen
concentration and water velocity on the mean dry weights of newly hatched
coho salmon fry developing from embryos reared throughout development at the
various combinations of oxygen concentration and water velocity represented
by the intersection of the curves. After Shumway, Warren, and Doudoroff
(1964).
80
-------
00
3.Or-
2.0
1.0
i
-1.0
Mean Temperature (°C)
O Control 9.4
A Intermediate I2£
D High 15.2
O
4.O
8.O
I2.O
I6.O
2O.O
FOOD CONSUMPTION RATE (% /day)
Figure 21. Relationships between mean rates of food consumption and growth of groups of steelhead
trout kept at different fluctuating temperatures in spring experiment. Rates of food consumption and
growth are expressed in dry weights. The initial mean wet wei,ghjL.of the fish was 2.29 g. Numbers in
parentheses indicate the number of fish which died in a treatment. After Wurtsbaugh (1973).
-------
/OO*-
oo
K)
I
\ 4O-
FECAL AND NITROGENOUS
WASTES (Aw)
SPECIFIC DYNAMIC ACTION
(Ad) AND ACTIVITY (Aa)
STANDARD METABOLISM (A- )
GROWTH (Ag )
-20 - •
i i I i I i
O 2O 4O 6O 8O IOO O 2O 4O 6O 8O O 2O 4O 6O
ENERGY IN CONSUMED FOOD (cal/kcat salmon/day)
Figure 22« Energy budgets showing relationships between food consumption rate, energy and material uses
and losses, and dissolved oxygen concentration for juvenile coho salmon in laboratory studies at 15C
in the summer. The influence of both food availability and oxygen concentration on scope for growth are
apparent in differences in growth parts of these budgets. After Thatcher (1975).
-------
b
14
13
12
//
10
8
C. ORYZAE
R. DOMINICA
1
1
I I
'
I
\
1
J L
16
2O 24 28 32
TEMPERATURE (°C)
36
4O
Figure 23. The finite rate of increase (X) as a function of temperature and moisture for the grain
beetles, Calandra oryzae and Rhizopertha dominica living in wheat. Rates of increase are indicated
by contour lines that describe conditions with identical values of \. After Andrewartha and
Birch (1954).
-------
033%
Exploitation Rate
1.0 FOOD LEVEL
1.5 FOOD LEVEL
20 25 30 35 40 45
Biomass (g)
Figure 24. Curves indicating relation of yield of guppies per 3-week brood
interval to biomass and exploitation rate (indicated percentages) at each
diet level. Points indicated for 0 percent exploitation rate are average
population levels for the 3 weeks immediately before exploitation. After
Silliman (1968).
84
-------
In Figure 25, we illustrate how different levels of productivity
determine the magnitude of production curves for trout in laboratory
stream systems. Production is defined as total tissue elaboration,
and it can be determined as the product of mean relative growth rate and
mean biomass. Different light levels and different current velocities led
to different algal community structures, different levels of food (insects)
availability, and so to different relative growth rate relationships with
biomass, and thus to different production curves. Such systems of produc-
tion curves entail not only differences in production performance under
different conditions but also differences in the capacities of the
laboratory stream systems to produce trout—the productivities of the
streams for trout.
Toxicants can affect production rates by the direct effects they
have on the growth of individuals in a population. Unstabilized effluents
of the kraft pulp and paper process can, at some concentrations, directly
affect the scope for growth performance of salmonids at different rates of
food consumption (Fig. 26A). This can be shown to lower production curves
for groups of salmonids, even when food availability in laboratory streams
remains unaffected (Fig0 26 B). Seim et al. (1977) have summarized some
of the research on this problem conducted at our laboratory. With
other toxic substances, for example dieldrin introduced at very low
concentrations, effects on the production of fish (Fig. 27B) appear
to be mediated more by reduction in the availability of their food
resources than by direct effects on food utilization for growth (Fig.
27A).
The dynamics of community structure—the changes in abundances of
constituent populations—whether measured in laboratory or natural communi-
ties, demands the use of theories or models, if it is to be understood in
terms of environmental factors and interactive individual organism and
population performances. This is so not only for dynamic but also for
steady-state community behavior. Even under constant environmental condi-
tions, population densities and community structure are time variant until
some possible steady state is leached (Fig. 12). But any steady state
that may be reached is only one of an infinite number of possible such
states (Fig. 12). And these, are determined by the levels of all
environmental factors including any effective toxicants present (Figs.
6, 12). One reason we earlier described the population system of algae
and protozoa developed by Taub (1973) was her synthesis of empirical labora-
tory research with simulation modeling. This permitted her to
determine nutrient concentration, light inentsity, and exchange flow rate
scopes for steady-state density structure, or performance, of the two-
species system she studied (Fig. 13). The introduction of an effective
toxicant would directly or indirectly affect the interacting populations
(Fig. 12), and so alter their dynamic and steady-state behavior, their
scopes for density performance, and scope for persistence of the two-
population system.
85
-------
Growth rate - biomass relationships
Production curves for trout in
laboratory streams having high
basic capacity to produce the
food organisms o1 trout
Production curve for trout in
laboratory stream having an
intermediate basic capacity
to produce the food organisms
of trout
Production curves for trout -
in laboratory streams hav-
ing low basic capacity to
produce the food organisms
of trout
500
450
400
350 O
CM
300
25O
20O
ISO
100
50
O
3
3
O
10 15 20 25 30 35
Trout biomass (kcal/m2) .
40
45
50
Figure 2S. Approximations of biomass-production curves for cutthroat
trout in the laboratory stream ecosystems having five different basic capaci-
ties to support trout production in the spring 1968 experiment. After Warren
(1971).
86
-------
30
23
20
o>
0)
o
I
10
-S
-10
E
-2"
c
o
o
a
T3
O
0.4
0.3
0.2
O.I
-O.I
-0.2
Effluent concentration
Volume BOO
O 0.0% 0.0 mg/l
O.S
2.0
3.0
20 40 60 80 (00 (20 140 160
Food consumption rate (mg/g/day)
180 200
Effluent concentration
Volume BOD
O 0.0 % 0.0 mg/l
A 1.5 3.0
0.2
0.4
0.6
O.8
1.0
1.2
1.6
Biomass (g/m2)
Figure 26. Relationships between food consumption and growth of juvenile
chinook salmon at different concentrations of kraft process pulp and paper
mill effluent, when the fish were fed controlled rations in aquaria (A). Aid
(B) relationships between the biomass of juvenile chinook salmon and their
production in laboratory streams receiving similar effluent and in streams
not receiving such effluent. After Tokar (1968) and Seim et al. (1977).
87
-------
o»
3
ft
o
I
o
Dieldrin
concentration
O 0.00 ppb
A 0.06
O 0,6
O.I 0.2 0,3 0.4 0.5
Insect density (g/mz)
0.6
0.7
6r
3
o
t>
£2
Q.
Dieldrin
concentration
O O.OO ppb
A O.06
D 0.6
10 IS
Biomass (g/m2)
25
Figure 27. Relationships between the growth rate of sculpins and the
density or biomass of the insects upon which they fed in laboratory streams
into which different concentrations of dieldrin were introduced (A). And (B)
relationships between the biomass of sculpins and their production in the
same streams. After Warren (1971),
88
-------
D. ON THE ADEQUACY OF EMPIRICAL INVESTIGATIONS IN ENVIRONMENTAL TOXICOLOGY
Before we finally move on to consider the entire matter of design,
interpretation, generalization, and application of theoretical and empirical
studies in environmental toxicology, let us briefly consider the general
problem of the adequacy of empirical investigation. For the problem of
empirical adequacy in environmental toxicology is the problem of empirical
adequacy in biological investigation confounded by the introduction of
innumerable toxic substances having practically unknown action and effect
sequences. Whatever is an empirical difficulty in biology must be a
.much greater empirical difficulty in environmental toxicology, and the
addition of human social dimensions further aggravates this difficulty.
Before considering criteria of empirical adequacy, we have first to
consider the matter of what it is that empirical work is to be adequate for :
What are its objectives, or the problems to be solved? For that part of
environmental toxicology with which we have here been concerned, empirical
investigations could have as their ultimate objectives: the screening of
toxic materials for decisions as to their possible introduction into the
environment; development of criteria for setting standards to protect
various organisms, including man, and even biological communities; explana-
tion and understanding of the behavior and effects of toxic substances in
ecological systems for levels from the individual organism upward; and
some combination of these. Some notion of prediction is generally involved
in application of empirical results or in testing of explanations. But
we, at least, are always uncomfortable about screening decisions and standards
based on little explanation and understanding, although we do recognize
the need for decisions and standards even when explanation and understand-
ing may not be entirely adequate.
Yet some sort of rationale, based on some sort of understanding of a
problem, must always be involved in human decisions. With regard to the
empirical basis of any such understanding, we would be well advised to
remember the raw data of experience are in no way an explanation, no
matter how extensive the data may be. The data present a need for expla-
nation: the data are the explanatory problem, not its solution. At most
from data we can hope to get empirical generalizations by some sort of
inductive process (Fig. 1). But many empirical generalizations are usually
necessary to account for vast accumulations of data. And it is difficult
to see the relationships among these generalizations and the boundary
conditions for application of any one of them. Moreover, empirical generali-
zations often seem to be contradictory, even though they might not be,
were we to understand more. Fegardless, the objective of smpiricial work
should be attainment of the most general and reliable empirical generaliza-
tions possible; not quantitative prediction, for which it is generally quite
unsuited.
But by what criteria are we to evaluate empirical studies in environ-
mental toxicology? In an earlier section, we suggested that theories
should, at least in part, be evaluated according to their dimensional,
dynamic, and empirical adequacy. Perhaps laboratory models of biological
89
-------
systems should be evaluated on the basis of their representational adequacy
and the generality of their results. In some sense, adequacy with regard
to all other pertinent empirical work, with regard to conceptual framework
considerations, and with regard to theoretical considerations must be
involved. No empirical result can be evaluated outside a broader context
of empirical experience, conceptual framework, and theory. Neither can it
be applied to solution qf problems in environmental toxicology without
some conceptual framework or theory consciously or subconsciously being
employed. It is for this reason that we emphasize that conceptual frame-
work and theoretical adequacy as well as representational adequacy and
generality ought to be employed as criteria for evaluation of empirical
work in environmental toxicology.
By representational adequacy, we mean that, in dimensional and
dynamic terms, a laboratory biological system should with some positive
degree of reliability represent the classes of objects and relations of
a natural system of interest. . Both the organismic system and the environ-
ment involved in the biological system in the laboratory should satisfy
this criterion. Because we have access to natural organismic systems for
laboratory studies at the individual organism level, and because reason-
ably representational environments can often be developed for individual
organisms, we believe representationally adequate laboratory studies of
individual organisms are possible. At the population level, for small
organisms with short life histories, some degree of representational
adequacy may be achieved, although environmental considerations become
more critical. But the problem of creating representationally adequate
multispecies systems in the laboratory is much greater. For, at this
level of organization, we are deeply involved in determining not only the
environmental part of the system but the very nature of the organismic
system itself: we in large part create the capacity of the multispecies
system by bringing together individual species having different capacities
of their own.
The criterion of generality of the results of laboratory biological
system studies is exceedingly difficult to achieve, or at least to know
that it has been achieved. And, yet, in some respects it is as important
as the criterion of representational adequacy. For no matter how well we
may have represented some particular biological system in nature, our
results can be of only very limited scientific and practical value if they
cannot be applied to other biological systems. This again raises the
whole problem of empirical generalization: we can at most study only a
relatively few biological systems, and yet our problems are with many; we
must generalize, but with the limited possibilities of empirical experience,
we cannot know the boundaries of applicability of empirical generalization.
This is one of the most fundamental reasons for theory as a scientific
deductive system through which empirical generalizations can be deduced,
explained, and at least theoretically bounded. Adequate schemes of
classification of organismic systems on all levels of biological organiza-
tion are absolutely essential to deal with the problems of generality.
This is as necessary for the conceptual and theoretical parts of our
investigative endeavors as it is for their empirical parts. Only
90
-------
through the development of more adequate classification schemes at all
levels of biological organization can we begin to deal with the problem
of generality of the results of our investigations in environmental
toxicology.
91
-------
V. CONCEPTUAL, THEORETICAL, AND EMPIRICAL EFFORTS TO
RESOLVE PROBLEMS IN ENVIRONMENTAL TOXICOLOGY
A. CONCEPTUAL AND THEORETICAL CONTEXTS OF PROBLEM PERCEPTION,
DEFINITION, AND UNDERSTANDING
t
We must now return to the matter of evaluating the objectives of
laboratory ecological system research in relation to problems in environ-
mental toxicology. In earlier asserting this need, we acknowledged that
such evaluation of objectives must be more subjective than the evaluation
of methodological approaches in relation to objectives. But we have found
that even the evaluation of methodologies is neither simple nor in itself
objective: There can be valid differences of opinion as to the adequacy
of particular methodological approaches to the achievement of the same
objectives, for presumably such differences of opinion always have some
rational basis for those holding them. Nevertheless, we cannot know the
basis of these differences without at least some articulation of their
presuppositions, either at the level of conceptual frameworks or at the
level of theories. Without some specification of conceptual frameworks
and theories, consideration of the adequacy of empirical approaches to
objectives can hardly be rational.
Evaluation of objectives of research in relation to apparent problems
is an even more difficult matter. This is so for several reasons. First,
there.can be valid disagreement as to the relevance of the objectives of
particular scientific work to problems perceived and even defined in the
same way. Second, problems associated with the same phenomena are not
always perceived as being the same. And finally, problems in environmental
toxicology are not simply scientific problems, perceived only by scientists
and to be approached only with scientific objectives and methodologies.
So, as with the evaluation of methodologies, the evaluation of the relevance
of objectives to problems in environmental toxicology can be no simple and
direct process, but rather one that must involve employment of conceptual
frameworks and theories. But at this level of problem perception, definition,
and understanding, conceptual frameworks and theories must involve that
which is sociological as well as that which is more simply natural science.
For evidence of great differences in perception of problems associated
with the same phenomena, we need look no further than the domain of
phenomena and problems stimulating preparation of this article. Publication
by Rachel Carson (1962) of her book Silent Spring, in which she
articulated her view of the potential hazards of pesticides to man and his
environment, stimulated a sometimes bitter controversy among scientists
and laymen alike. The correctness or incorrectness of Miss Carson's
view and whatever rational basis it may have had is not here at issue.
What is at issue is that apparently deeply held beliefs of biochemists and
ecologists and of agriculturalists and environmentalists led to such
different perceptions of whatever problem or problems may have existed
that the controversy over this book became extremely bitter. Neither
is the correctness of the many very different perceptions of this problem
that arose at issue here. Such differences in views may not in themselves
be undesirable; indeed, they may be desirable, for no single view of
92
-------
complex phenomena and problems is likely to provide for understanding and
problem solution. But without articulation of the different conceptual
frameworks from which contrary perceptions of the problem arose and with
no relevant theory, consideration of differences in problem perception and
of what the total problem might really be like could not be rational:
indeed, it became very irrational on all sides. Remnants of this contro-
versy persist to this very day, biochemists, ecologists, agriculturalists,
foresters, and environmentalists continuing to disagree in the absence of
adequate conceptual frameworks and theories. And because toxic substances
in the environment present a social problem of vast complexity and of
potential as well as actual importance, the sort of problem science as
presently pursued cannot handle, it has been left to legislative, judicial,
and administrative bodies for resolution, still with inadequate scientific
understanding available.
Thus it would seem, in the whole domain of environmental toxicology,
we have not only to conceive of and employ conceptual frameworks and theories
in the evaluation of objectives in relation to problems, we have also to
employ these in attempting to perceive and define just what the problems may
be. But these could be no ordinary conceptual frameworks and theories, for
they would need to subsume that which is social and technological as well
as that which is more nearly the domain of the natural sciences. Less
than this is apt to ever leave problem perception, definition, and solution
short of being rational. With realization of our ever too limited social
and scientific capacities, even vague perception of such a need may cause
us to back away and do other things now easily rationalized but probably
ever too partial to solve our problems. And yet, surely the continued
need to employ toxic substances—pesticidal, medicinal, and other—and the
continued potential hazard of these is of such vast importance to man
that he will expend great resources and efforts in problem evaluation.
Some of these resources and efforts ought to be expended in approaching
that which may seem unapproachable but which, so long as the possibility
of some success exists, cannot justly be avoided. Such frameworks deserve
and need to be articulated to provide greater rationality in human social
and scientific endeavors. That the domain of environmental toxicology is
vast and may never be subsumed with full adequacy in no way lessens the
need for us to begin to approach the development of conceptual frameworks
and theories sufficiently universal for adequate problem perception,
definition, and understanding.
In the recorded history of man, so often his ideas and other creations
have not accomplished that for which they were intended and have led to
social problems, we cannot help but wonder what creations ought to be
encouraged and what might best be avoided. But it would not be desirable,
and certainly would not in the long run be possible, to place any controls
whatsoever on human imagination and the search for understanding and
better world systems for man. We do not believe that understanding or the
ideas for concrete creations can be shown to be responsible for social
problems but rather that problems are owing always to insufficient under-
standing and the uses to which ideas are put. In this, our harmful
experiences with toxic substances are no exception. If there had been or
93
-------
could be something like adequate understanding of social and natural
systems in all their interrelationships, and were individuals and societies
to behave more rationally, these and other human experiences could have
been quite different, -much more beneficial and much less harmful. But in
the absence of universally articulated understanding of social-natural
systems, their capacities and possible behaviors under all possible
abstractly conceivable conditions, we have no reason to suppose but that
too many human experiences will continue to be harmful. It is not the
idea or invention of a terribly persistent and toxic substance but rather
the decision to develop and employ it widely that can lead to economic and
social dependencies, very harmful side effects, and later attempts to
correct apparent errors, these attempts then leading to further social
problems. If there could be a general theory of management of toxic
substances, which would universally articulate understanding of natural
and social systems, there might be some hope that the long sequence of
human errors could in the future be more often interrupted. But it seems
almost impossible to conceive of what such a theory might be like—what
its form and content would be.
The problem of management of toxic substances is not unlike the
problem of management of natural resources. Even were we to have a very
universal theory of, say, fisheries or forest exploitation—which we do
not—we would yet not have a theory of management of these resources.
A universal theory of fisheries or forest exploitation could tell us the
yields obtained at different rates of exploitation under various physical
and biological conditions and what other ecosystem effects would occur.
But it would not tell us what a society might choose to do with these
.resources or want of these ecosystems under various circumstances deter-
mining the behavior of that society. Failures of present day fisheries
and forest management to account adequately for societal demands have led
to entry by legislative and judicial bodies into the business of resource
management, usually to the chagrin of the professionals. Because toxic
substances have been employed widely in attempts to manage renewable
resources and have thus led to some of our most prevalent problems, it_is
not inappropriate, for the moment, to consider these management problems
together. Indeed, a theory for one would need to entail a theory for the
other, in many applications of the theory.
In some sense, such a theory would need to subsume all the biological
and physical effects of different kinds and levels of use and all possible
societal behaviors leading to or influenced by these possible uses. Only
in this way could the many possible interrelated action and effect webs be
explained, understood, and anticipated to the extent necessary to avoid
resource, toxic substance, and social problems. We are thinking here of
theory as a scientific deductive system, articulated in abstract language,
interpretable in observational language as necessary, and universal in the
sense of being invariant and general. We recognize that most if not all
our colleagues in the natural sciences and friends in the social sciences
will say no such theory is possible, certainly not now, probably not ever.
And they are probably correct. But somehow the question keeps returning
'to haunt us: Can we afford to avoid the attempt?
94
-------
Any such attempt, it seems to us, should begin not at the conceptual
level of formal theory but rather at the level of what we have been calling
conceptual frameworks. The rush to formality, in our opinion, has too
often in biology and the social sciences led to excessively partial views
that can only be externally inadequate, whatever the merits of the
achieved internal consistency.
Theories arise out of implicit or explicit conceptual frameworks,
which give theories much of their meaning. And at this stage it is not
the language or the form of theory that must concern us. Rather, we must
first achieve some outermost abstract description of what the domain of
concern, here social-natural systems, might be like. This is what Warren,
Allen, and Haefner (MS) attempted to do for biology with the set of four
generalizations presented very early in this article. Such an abstract
verbal description, with the aid of its rules, should tell us something of
what a theory or theories would need to do to articulate adequately
that which lies behind, determines, and so could explain phenomena in the
domain of concern. Only then can we rationally approach the matter of
language and form of theories to be articulated.
t Communication with full understanding even within broad disciplines
ficult. Between different disciplines, say the physical, the
biological, and the social sciences, it is not at all apparent that
communication occurs sufficiently well for the integrated thought and
understanding necessary to achieve unified theory. Among the most important
reasons for making conceptual frameworks explicit is that, even when
different views are held, rational discussion based on the frameworks can
proceed, for the basis of arguments and judgments can be examined by all.
And it is possible that a universally adequate conceptual framework for
biology, differently interpreted, would be found to be in some degree
adequate for the social sciences. If it were found not to be so, .it
might, for the same underlying reasons, not be so adequate as it was
believed to be for biology. In any event, it is quite possible that some
conceptual framework could be developed that would subsume the biological
and social sciences. And, it is our belief, that even before adequate
theoretical languages and forms could be found and employed in the articu-
lation of any sort of general theory of management of toxic substances,
the conceptual framework itself would be found to be useful in the reso-
lution of some problems. Indeed, the process of developing a framework,
however successful the product, would be beneficial in bringing the differ-
ent disciplines together, for many of our problems appear to have their
roots in our failure to do this.
In the absence of articulated conceptual frameworks and theories
adequate to subsume that which is social as well as that which is more
directly physical and biological, we cannot rationalize problem perception,
problem definition, the relevance of objectives to problem solution, and
approaches to these objectives in environmental toxicology. But anyone
thin&Lng for awhile about the whole domain of environmental toxicology
can delineate what appear to be problems, possible social and scientific
objectives perhaps relevant to the solution of these problems, and approaches
95
-------
to these objectives. That this can be done is in itself evidence of some
a p;?iori conceptual basis for any such effort and delineation. Never-
theless, without articulation of this conceptual basis, either as a frame-
work or a theory, we cannot demonstrate whatever rationality our delineation
may have; although, in view of rather obvious human social problems, the
effort at least would appear to be pragmatically rational. In any case,
we can here not proceed without some delineation of social problems and
institutional objectives and approaches—as we have attempted to do in
Figures 28 and 29—before we proceed to evaluate the relevance of objec-
tives of laboratory ecosystem research in environmental toxicology.
In delineating problems, objectives, and approaches in environmental
toxicology, we ought to begin with some theoretical notion of the relations
of societies to their total environments. Such a notion, however, would
be exceedingly difficult to articulate, because man's ideas become
part: of the environment of particular societies and influence their behaviors
as wholes: To think of the environment of a society only in mere physical
terms is to leave out much that would be necessary to understanding of its
behavior. Now all organismic systems, at whatever levels of organization,
beha.ve as though they have goals. However we may account for such
behavior in nonhuman organismic systems, man certainly perceives goals,
and the ways in which these are perceived have much to do with the behaviors
of human social systems. Goals and the possibilities of achieving them
are determined by geographic, cultural, political, technological, and
scientific heritages and states, as we suggest in Figure 28. Social prob-
lems are perceived in the creation and utilization of human environments
intended to satisfy needs and aspirations—problems that perhaps can
generally be categorized as being resource, cultural, political, or
technological (Fig. 28). With the intent to resolve these problems,
implicit and explicit objectives are imperfectly perceived and articulated,
and institutional approaches to these problems and objectives are conceived.
But these objectives are not always externally consistent and their pursuit
may aggravate and create problems as well as lead to some solutions (Fig.
28). And the methods of human institutions may or may not, singly or
together, be adequate for realization of objectives.
We may think of the institutional approaches to social problems as
being legislative, judicial, administrative, educational, technological,
and scientific (Fig. 28). Here we are most directly concerned with scien-
tific approaches, a.nd Figure 29 is our cursory attempt to delineate some-
thing of the goals, problems, objectives, and methods of the scientific
institution. .But, so long as our ultimate concern is with social goals,
problems, and objectives, as it must be in environmental toxicology, our
scientific institution must be seen to contribute to social achievement in
the context of all human institutions. We can take the ultimate goal of
man's scientific institution to be the continuous advance of human under-
standing (Fige 29). Such understanding as is achieved may be used in the
resolution of social problems.
Understanding is advanced by theoretical and empirical explanations
within the context of existing conceptual frameworks. And scientific
96
-------
JHUMAN ENVIRONMENT
MAN
SOCIETAL, GROUP, AND INDIVIDUAL
GOALS - POSSIBILITIES
ACCORDING TO GEOGRAPHIC, CULTURAL, POLITICAL
TECHNOLOGICAL, AND SCIENTIFIC HERITAGES AND STATES
SOCIAL PROBLEMS ARE PERCEIVED
IN THE CREATION AND UTILIZATION OF HUMAN ENVIRONMENTS
; INTENDED TO SATISFY NEEDS AND ASPIRATIONS
RESOURCE CULTURAL POLITICAL TECHNOLOGICAL
PROBLEMS PROBLEMS PROBLEMS PROBLEMS
IMPLICIT AND EXPLICIT OBJECTIVES FOR RESOLUTION OF SOCIAL PROBLEMS
IMPERFECTLY PERCEIVED AND ARTICULATED
INSTITUTIONAL APPROACHES TO SOCIAL PROBLEMS ARE CONCEIVED
LEGISLATIVE JUDICIAL ADMINISTRATIVE EDUCATIONAL TECHNOLOGICAL SCIENTIFIC
OBJECTIVES OF INSTITUTIONAL APPROACHES
INTENDED TO RESOLVE SOCIAL PROBLEMS
BUT NOT ALWAYS CONSISTENT AND THEIR PURSUIT
MAY AGGRAVATE AND CREATE PROBLEMS
METHODS OF HUMAN INSTITUTIONS
MAY OR MAY NOT, SINGLY OR TOGETHER,
BE ADEQUATE FOR REALIZATION OF OBJECTIVES
objectives and methods of scientific, technological, educational, and governmental
institutions generally involve redefinition of social problems in ways making
institutional approach possible, but such redefinition often makes
approaches ineffective or irrelevant
Figure 28. Delineation of human goals, problems, institutions, and
institutional objectiv.es and approaches.
97
-------
HUMAN ENVIRONMENT
MAN'S SCIENTIFIC INSTITUTION
GOAL
THE CONTINUOUS ADVANCE OF HUMAN UNDERSTANDING
SUCH UNDERSTANDING AS IS ACHIEVED MAY BE
USED IN THE RESOLUTION OF SOCIAL PROBLEMS
BUT MAY ALSO CREATE NEW SOCIAL PROBLEMS
UNDERSTANDING IS ADVANCED BY
THEORETICAL AND EMPIRICAL EXPLANATIONS WITHIN
THE CONTEXT OF EXISTING CONCEPTUAL FRAMEWORKS
PERCEPTION AND DEFINITION OF SCIENTIFIC PROBLEMS
THEORETICALLY OR EMPIRICALLY
PROBLEMS THAT ARE FIRST PERCEIVED AS
BEING SOCIAL AND GENERALLY REDEFINED SO
AS TO MAKE SCIENTIFIC APPROACH POSSIBLE,
BUT IN SO DOING, APPROACH MAY BE MADE
INEFFECTIVE OR IRRELEVANT
SCIENTIFIC OBJECTIVES
PRESUMED TO BE RELEVANT TO
SOLUTION OF DEFINED PROBLEMS
SCIENTIFIC METHODS
PRESUMED TO B£ ADEQUATE
FOR REALIZATION OF OBJECTIVES
THEORIES
PROBLEM PERCEPTION AND DEFINITION
FORMALLY EXPLAIN OR GIVE MEANING
TO OBSERVATION AND EXPERIENCE -
THEMSELVES GIVEN MEANING NOT ONLY
BY OBSERVATIONAL VALIDATION BUT
ALSO BY PREVAILING CONCEPTUAL
FRAMEWORK
MODELS OBSERVATIONS
PROBLEM PERCEPTION AND DEFINITION -
MUST BE EXPLAINED OR GIVEN MEANING
BY CONCEPTUAL CONSTRUCTS
Figure 29. Delineation of the goals, problems, objectives, and approaches
of man's scientific institution.
98
-------
problems are defined, either theoretically or empirically, within the same
contexts. In this process, we would note that problems which are first
perceived and defined in social terms are generally redefined so as to
make the scientific approach possible. But such redefinition, as neces-
sary as it may be in the pursuit of science, often makes the scientific
effort ineffective or even irrelevant to possible solution of the problems
originally motivating the effort. We are reminded that Kenneth Boulding
once observed science to be the process of .substituting unimportant
problems we can solve for important ones we cannot.
B. APPARENT AND POSSIBLE OBJECTIVES OF LABORATORY ECOLOGICAL SYSTEM STUDIES
We have prepared Figure 30, as much for ourselves as for others, to
help keep the objectives of laboratory ecological system studies in the
context of all scientific problems, other biological problems, and other
approaches in environmental toxicology. As important as may be the
possible objectives and results of laboratory ecological system studies,
they directly relate to only a part of the scientific problem, let alone
technological, governmental, and educational problems, in environmental
toxicology. And the possible contribution of such laboratory system
research is better seen in relation to the whole.
In defining apparent and possible objectives of laboratory ecological
system studies in environmental toxicology, we have attempted to be clear
in distinguishing among them and to be reasonably comprehensive, without
excessive proliferation. Of course, decompositions into subobjectives
of the major objectives that we do present is desirable and is done in
practice. Although we have already given examples from the literature
of work that appears to have been directed toward these objectives, we
will make no attempt here to classify this work according to objectives.
Any such classification, we believe, will be evident enough.
In Figure 31 we list, in the first line, our view of the major
apparent and possible objectives of laboratory ecological system studies
in environmental toxicology: assay of the relative toxicity of substances;
determination of the behavior of toxic substances; determination of the
effects of toxic substances on individual organisms; determination of the
effects of toxic substances on biological populations; and determination
of the effects of toxic substances on biological communities. To be more
general, these objectives should probably have included the notion of
cause as well as effect. For, certainly, it may reasonably be argued
that one of the reasons for laboratory studies is to make possible causal
or operational explanation of phenomena we perceive in nature. But causal
relationships cannot be observed, as we earlier emphasized. And so far as
ecological explanation is involved, adequate observation of effects on
performances at individual, population, and community levels of biological
organization can be articulated into operational or causal explanations,
but only with dimensionally, dynamically, and empirically adequate conceptual
structures that are causal-deterministic.
99
-------
PROBLEMS IN ENVIRONMENTAL TOXICOLOGY
SCIENTIFIC PROBLEMS
(OTHER PROBLEMS TECHNOLOGICAL. GOVERNMENTAL, AND EDUCATIONAL)
(Problem in
Social Sciences)
Problems in
Biological Sciences
(Problems in
Physical Sciences)
(Problems of uirect Effect's
of Toxic Substances on Hunan Health)
Problems of Human Health and
Veil Being Resulting from
Introduction of Toxic Substance*
into Mm'» Biological Environment
Accumulation of Toxic
Substanceii in Biological
Products Utilized by Man
Reduction in Production
of Biological Resources
Caused by Toxic Substances
Toxic Alteration of
the Structure, Persiitence,
and other Performances of
Ecological Systems of
Importance to Man
(Toxic Problems at Suborganinic
o:r Physiological Levels of Organization)
Toxic Probleu at Ecological
Levels of Organization
Transport and Accumulation
of Toxic Substances
in Biological Coneunities
Toxic ALter*ticn of
Survival, Development,
Growth, Reproduction,
and Behavior of
Individual Organise*
Toxic Alteration of
the Structure, Persistence. |
and other Performances of
Biological Populations and ,
Communities j
OBJECTIVES OF ECOLOGICAL STUDIES OP ENVIRONMENTAL
TOXICOLOGY DERIVE FROM PROBLEMS SUCH AS ABOVE
(Objectives of Studies of
Natural Ecological System
Objectives of Studies of
Laboratory Ecologic*! Systems
r~
I,,
Evaluation of the
«
-------
ASSAY OF THE
RELATIVE TOXIC ITY
OF SUBSTANCES
DETERMINATION OF
THC BEHAVIOR OF
TOXIC SUISTANCES
DETEIMINATION OF THE
EFFECTS OF TOXIC SUfSTAHCES
ON INDIVIDUAL ORGANISMS
DETERMINATION OF THE
EFFECTS OF TOXIC SUBSTANCES
ON BIOLOGICAL POPULATIONS
DETERMINATION Of THE
EFFECTS OF TOXIC
SUBSTANCES ON 110-
LOG 1CAt COMMUNITIES
OTHER SPfCIEt
AND DEGRADATION OF TOXIC
SUBSTANCES IN MULTISPECIES
SYSTEMS
VALUATION OF
CAPACITIES ANO
PCRFOAMANCES OF
INDIVIDUAL OR«AN-
ISMS UNDER TOXIC
AND OTHER CONDITIONS
MHEN NOT INTERACTING
j*ITH OTHER SPECIES_J
(VALUATION OF
CAPACITIES AND
PER ORMANCES OF
INO VIDUAl ORGANISMS
UNDER TOXIC AND
EVALUATION OF RELATIVE
ffBKSSSffS
N OF DETERMINATION OF
1
'
EVALUATION OF
CAPACITIES AND
PERFORMANCES OF
POPULATIONS UNDER
TOXIC AND OTHER
CONDITIONS HHEH
HOT INTERACTING
MITH OTHER
SPECIES
EVALUATION Of
CAPACITIES AND
PERFORMANCES OF
POPULATIONS UNDER
TOXIC AND OTHER
CONDITIONS WHEN
INTERACTING MITH
OTHER SPECIES
EVALUATION OF I
CAPACITIES AMD I
PERFORMANCES OF
MULTISPECIES SYSTEMS
UNDER TOXIC AND
OTHER CONDITIONS
CAPACITIES TO: PERFORMANCES Of: !
PERSIST PERSISTENCE
INCREASE NUMBERS NUMERICAL INCREASE
ELABORATE TISSUE PRODUCTION
[CAPACITIES TO:
PERSIST
DEVELOP
PROVIDE FOR
POPULATIONS
OF TOX.C SUISTANCES
«poDUCTION
PERFORMANCES OF:
PERSISTAMCC
DEVELOPMENT
STRUCTURE
POPULATION
SUCCESS
=2
o
m
3
=
I
03
Figure 31. Apparent and possible objectives of laboratory ecosystem studies in environmental toxicology.
-------
Major objectives, such as the five we have just mentioned (Fig. 31),
can be approached methodologically in the field and in .the laboratory, as
well as theoretically, in various ways. But here we are concerned most
immediately with laboratory studies, especially ecological system studies
in environmental toxicology. One point that we have endeavored to make
clear in the second line of objectives in Figure 31 is that laboratory
studies of individual organisms and populations not interacting with
other species, that is, not parts of laboratory ecosystems, are an important
way of approaching the relevant major objectives. Indeed, they may be a
much better way than poorly designed laboratory ecosystem. We have, for
this reason, shown enclosed in brackets in the second line of Figure 31
subobjectives for studies of individual organisms and populations not
interacting with other species.
Achievement of the objectives not enclosed in brackets in the
second line of Figure 31 would require investigation of multispecies
systems or, at the individual and population levels, at least the presence
of other species. The first of these objectives, evaluation of the relative
toxicity of substances in multispecies systems, relates essentially to the
employment of a complex organismic system as a reagent for assay of the
toxicity of materials, much as individual organisms are employed in
bioassays. The second objective in the second line of Figure 31 is
determination of transport, distribution, and degradation of toxic sub-
stance in multispecies systems. This appears to be the objective of most
of the recent laboratory ecosystem studies in environmental toxicology.
The three remaining objectives not enclosed in brackets in the second
line of Figure 31 relate to determination of the effects of toxic sub-
stances on individual organisms, populations, and multispecies systems.
And, in these objectives, we make what is probably the most important
distinction of all., the distinction between the capacities and the per-
formances of organismic systems. This distinction has rarely been made
in either theoretical or empirical biological studies. And, yet, without
it there is no way that we can conceive for achieving the sorts of
universal explanations necessary for adequate solution of biological
problems, including those in environmental toxicology.
Capacity of an organismic system, as we earlier explained, is a
theoretical concept necessary to explain all possible performances of an
organismic system in all possible environments. But if we must know some-
thing of the capacities, not just the performances, of organismic systems
to be able to deal adequately with problems, in environmental toxicology,
and if capacity is a theoretical concept not empirically determinable, how
are we to proceed? We cannot directly and fully evaluate theoretical con-
cepts such as capacity, but indirectly and partially we can do so, or they
would be of no value to us. We can measure particular kinds of performances
of an organismic system under a wide range of combinations of environmental
conditions of interest, in the laboratory. Under each set of conditions,
the organismic system is likely to have a different level of any given kind
of performance. We can thus determine the effects of environmental
102
-------
conditions on the scope for performance of the organismic system of interest,
as we illustrated in Figures 13, 16, 17, 19, 20, 23, 25, 26, and 27. But
scope for performance is yet an empirical concept: it is only a partial and
indirect evaluation of the capacity of the system. For were the organismic
system of interest to have been tested under other sets of environmental
conditions, still other scopes for performance could have been determined
all within the capacity of this system, all likely to be of some natural
importance, and thus all of at least theoretical interest in the solution
of problems of environmental toxicology. And were our organismic system
of interest—individual organism, population, or multispecies system—to
have developed under different conditions, its scope for performance under
even the same subsequent conditions would very likely have been different.
From some potential capacity, the realized capacity an organismic system
develops will depend on environmental conditions during system development
(Fig. 3). Developmental effects of toxicants on the capacities of organis-
mic systems at all levels of biological organization should be of great
concern. We can never fully or directly know these, but we must get as near
to knowing them theoretically and empirically as we possibly can. Funda-
mentally, these are the reasons that we judge the distinction between
capac&y and performance, which we employ in articulating the objectives
given in Figure 31, to be most important.
Thus the last three objectives that are not bracketed in the second
line of Figure 31 relate to evaluation of effects of toxic substances on the
capacities as well as on the performances of individual organisms and
populations in multispecies systems and on the capacities and performances
of multispecies systems as wholes. In the third line of Figure 31, we
list some possible capacities and performances for each of these levels of
biological organization. Individual organisms have capacities to utilize
resources, survive, develop, grow, and reproduce. Evidences of these are
the performances: resource acquisition, survival, development, growth,
and reproduction, which we can measure. Similarly, population and multi-
species systems have some capacity to persist, which we can partially
evaluate by measurements of persistence, a performance. And, of course,
there are other capacities and performances, at all levels of biological
organization. In relation to the objective to determine transport,
distribution, and degradation of toxic substances in multispecies systems,
we should perhaps make one additional comment. Such behavior is, of
course, determined in part by the physical-chemical nature of the toxicant,
as is certainly recognized by all investigators. But what is equally
important and is not so generally recognized is that a measured
behavior of transport, distribution, and degradation is also a performance
of the biological system in which the behavior was measured. It is,
however, only one possible performance—determined by environmental
conditions and the state of the system—of all possible performances for
which that system has the capacity. Theoretical and empirical evaluations
of the capacity of a system provide the most general results we can hope
to obtain for that system.
103
-------
C. DESIGN OF LABORATORY ECOLOGICAL SYSTEM STUDIES IN
ENVIRONMENTAL TOXICOLOGY
The immediate explanatory objective of empirical studies is to obtain
sets of empirical generalizations (Fig. 1) relating to the problem and the
overall defined objectives motivating the work. In the present context,
the overall objectives pertain to the behavior and effects of toxic sub-
stances in ecological systems, much as we defined such objectives in the
previous section. We say the immediate explanatory objective of empirical
studies is to obtain sets of empirical generalisations because, in view
of the uniqueness of capacities and environments of organismic systems,
the domains of explanatory application of a particular empirical generali-
zation based on studies of a particular ecological system cannot be large.
Its application outside that domain is likely to lead to erroneous
decisions. Against this probability we can protect ourselves at the
empirical level of understanding only by the achieving of sets of empirical
generalizations based on studies of organismic systems having different
capacities and evaluated over many different sets of environmental condi-
tions. What each such ecological system in some sense represents and how
it is different from others being studied should be determined by design
and results. And the design of each should make possible the elucidation
of fundamentally important relationships that we can expect to exist in
systems at the level of organization it is intended to represent.
The design of a laboratory ecological system amounts to the selecting
or composing of an organismic system having capacities for performances of
interest and to the provision of suitable and generally interesting sets of
environmental conditions. When we intend to study an organismic system at
the individual or population levels of organization, the problem becomes
mainly one of providing suitable and generally interesting environmental
conditions. This is because the potential capacity of an individual organism
or a population o£ a given species has largely been determined by nature.
Even so, we must remember that, through developmental effects of the pro-
vided environment, the realized capacities of individual organisms are
in part determined by our design of laboratory ecological systems. But at the
multispecies level of organization we determine not only the sets of
environmental conditions but also the capacity of the system as a whole, by
the species of organisms we choose to include in a laboratory ecological
system. All of this we have endeavored to make clear in previous sections
of this article.
Still, with all the difficulties we have considered at length, Tiow are we
to design laboratory ecological systems so as to be able to induce from
observational data sets of empirical generalizations having explanatory
applicability over domains sufficient to be of much value in environmental
toxicology? First of all, empirical generalizations are based on general
concepts of objects and relations implicitly or explicitly derived from
classification schemes. To this we must direct much more attention.
The organismic objects of concern are individual organisms, populations,
and multispecies systems. For individual organisms and populations, the
taxonomic classification of species is important, in that it in some
104
-------
degree reflects evolutionary similarities and differences. Classification
into trophic types—autotrophic plants, herbivores, omnivores, carnivores,
and decomposers—is also important, and it is necessary if we are to deal
with multispecies systems. Classification of relationships among popu-
lations in multispecies systems is necessary if we are to know the domains
of empirical generalizations which state general relationships between
concepts of classes of objects. Examples of such relationships among
organismic objects are predation, parasitism, competition, commensalism,
and mutualism. Classification of physical environmental objects, rela-
tionships, and modes of operation is also necessary. F. E. J. Fry's (1947)
scheme of operational classification of environmental factors—lethal,
controlling, limiting, stressing, masking—should prove very useful at all
levels of biological organization, especially since he developed this scheme
to deal with general determination of the effects of environmental factors
on the scopes for performance of organisms. And, if we are to achieve sets
of empirical generalizations useful in environmental toxicology, we must
develop adequate schemes of classification of toxicants according to their
structure, behavior, modes of action, and effects. From classes of organis-
mic and environmental objects and relations, we can choose representative
ones 'to incorporate in our designed laboratory ecological systems ariti thus
simplify the problem of empirical generalization and clarify the domains of
explanatory applicability of individual and sets of empirical generaliza-
tions.
For laboratory ecological systems at the individual organism level of
organization, we should select species as broadly representative of taxonomic
and trophic classes as possible. And examples of a rather wide variety of
these classes should be investigated. And always we should evaluate the
capacities of the organisms we study by determining their scope for per-
formance under many different sets of environmental conditions, including
toxicants of different classes. Food or other energy and material avail-
ability, temperature, water availability, and other important factors in the
natural environments of the species to be studied should be varied over
interesting ranges in determining toxicant behavior and effects and scopes
for performance of individual organisms. Even if it is not our intention
to study multispecies systems, in the restricted sense we define these, it
may still be worthwhile to include in the environment of the individual
organism of interest individuals of other species that may alter its scope
for performance. Finally, our definition of an individual organism., as
an organismic system, incorporates the entire life history trajectory
of the individual, from zygote through development to and including
reproduction. Our empirical generalizations, as they apply to individual
organisms, must relate sets of environmental conditions to sets of
performances of different kinds—scopes for performance—at all life
history stages. In this way, generalizations pertaining to the individual
organism level will relate more nearly to generalizations pertaining to the
population level of organization than they would were we to deal only
with^particular life history stages.
105
-------
This is because a population differs from a simple aggregation of
individual organisms of the same species in that it has'the capacity to
persist through many generations of individuals as a result of life history
completion and reproduction of these individuals. Our empirical
generalizatins about populations should relate effects of particular classes
of toxicants to the persistence and other capacities of populations under
a wide range of other environmental conditions, through determination of
population scopes for performance. To do this, we must provide those
environmental conditions necessary for completion of individual life
histories and for the development and persistence of the resulting popu-
lation. Energy and material resource levels must be adequate, but should be
studied at different quantitative levels, and qualitative difference should
also be tested. And, as with the individual organism, environmental factors
such as temperature and water availability should be included in our evalua-
tions of population scope for performance. In addition, without creating
persistent multispecies systems, the capacities of populations can be fur-
ther evaluated by including in the environment of the population different
densities of the species such as competitors or predators.
Simply because it may sometimes be possible and useful to study the
capacities and performances of individual organisms and populations inter-
acting with other species that do not together form a system does not mean
that this is always the best approach, There are good reasons for evaluating
capacities and performances of individual organisms and populations in what
are indeed multispecies•systems having capacities for persistence and other
performances that the organisms and their populations separately do not
possess. And, of course, if we are interested in the behavior and effects
of toxic substances in multispecies systems, then we have no choice but to
study such systems, either in nature or in the laboratory. The creation of
multispecies systems in the laboratory—systems having their own, level-
specific, capacities to persist, develop, and provide for their population
and individual organism parts—requires considerable knowledge of the
biological characteristics and requirements of species considered for inclusion.
Moreover, unless success in developing a multispecies system is to be achieved
only by chance through an indeterminate number of trials, there must be some
theoretical means of deducing from life history characteristics and parameters
of candidate species the likelihood of our being able to form with them a
multispecies system having the capacity to persist through many generations.
To develop a laboratory ecological system that is indeed a multispecies
system, we must provide for completion of the life histories of individuals
of the species included. But, for the system to have the capacity to
persist, this alone is not enough. Multispecies system persistence is pos-
sible only if the capacities of the individual populations for numerical
and biomass increase are great enough at each step in the energy and
material transfer pathways to sustain successive steps. When this condition
has not been met, multispecies systems have not been created and toxic effects
of substances on a multispecies system as a whole and the behavior of the
toxicant in the system cannot be reliably evaluated.
106
-------
If we have indeed created a multispecies system having capacities for
persistence and organization, we ought to evaluate that capacity under
different sets of environmental conditions—we ought to determine the
scope for performance of that system under a range of conditions of
interest. This is because no particular set of environmental conditions
is of general interest, and the multispecies system will have different
performances, including structure, with changes in environmental conditions.
We should also note that laboratory multispecies systems are severely
limited in their capacity for development, or change in structure, because
"colonization" by new species is generally eliminated or severely restricted
(Warren and Davis, 1971).
Finally, however fully we may be able to evaluate it, any multi-
species system has its own capacity. The capacity of a multispecies system
in the laboratory we determine by bringing together individual organisms
of different species having their own capacities. Whatever may be this
capacity for performances—including transport, biomagnification, and
biodegradation of toxic substances—we have determined it by chance or by
design. It is true that each of the individual organisms have certain
capacities of assimilating, accumulating, and degrading toxic substances,
and these capacities nature determined. But nature does not determine the
capacity of any laboratory multispecies system as a whole. We ought to
design laboratory multispecies systems to include representative populations
and relations, including predation, competition, commensalism, and mutualism,
if these systems are to behave at all like natural ones. Even then, we
must continue to be aware that the capacity of any designed system is
peculiar to it, and general application of laboratory results to nature
should be made with great care and with the assumptions made explicit.
An empirical generalization, at the multispecies level of ecological
system, should make explicit the nature of the system from which it was
derived. And we need such generalizations for different multispecies
systems designed to include various combinations of species populations and
interrelations.
D. APPLICATION OF LABORATORY ECOLOGICAL SYSTEM STUDIES
IN ENVIRONMENTAL TOXICOLOGY
The goal of science—the increase in human understanding—ever must
be not only relevant to and concordant with other goals of society but
indeed must be one of the major goals of humanity, for creativity and the
search for understanding most uniquely characterize mankind. It is not
understanding but rather misuse or insufficiency of understanding that
leads to human social problems. Objectives of particular scientific
endeavors, even if not so relevant to specified social problems as might
be supposed, ought to be formulated in ways making their pursuit likely to
contribute to understanding and so to society. Redefinition of either
social or scientific problems so as to make possible their solution is not
simply an artifact of science: Defining problems in ways that the world
can be understood is the essence of science. We must not take too narrow
a view of what is or is not relevant to any human concern, including
problems in the domain of environmental toxicology.
107
-------
Human capacities may demand some narrowness of definition of social
and scientific problems and objectives for any single' institutional means
of approach to be employed. But we must endeavor to avoid the narrowness
of view that vitiates our efforts. Society through its scientific and
other institutions may at any time seem to demand of us'narrow views and
approaches intended to solve social and scientific problems. Perhaps this
is inevitable and even apparently necessary in the short run, but its
social and scientific consequences have too often been found to be incon-
sequential in the long run. Social and natural scientists owe it to
themselves and to society to take.broad views, and they have the respon-
sibility of teaching the necessity of this.
The objectives of laboratory ecological system research, as they
appear at the bottom of Figure 30 and as articulated in Figure 31, appear
to be relevant to scientific problems in environmental toxicology. And
their achievement would presumably be useful in solutions of technological,
governmental, and educational problems also necessary to solution of the
social problems facing us. Beginning at the top of Figure 30, we have
attempted to delineate, through successive decomposition of problems
finally to problems at ecological levels of organization, a pathway through
which the relevance of the objectives of laboratory ecological system
research might be informally evaluated. Even though this delineation is
in no way comprehensive, it does make apparent that most of the objectives
of laboratpry ecological system research can be approached in other ways
and that their realization, by whatever theoretical or empirical means,
can only very partially resolve pertinent problems in the biological
sciences and must be accompanied by solution of problems in the social and
physical sciences in order to contribute importantly to solution of
technological, governmental, and educational problems and the social prob-
lem itself. This conclusion must be obvious to all concerned, but never-
theless should be noted in a consideration of the relevance of the objec-
tives of any scientific approach to complex social problems such as those
occurring in the domain of environmental toxicology.
Our problem lies not with such superficial relevance of defined
objectives but with how even defined objectives are perceived and under-
stood differently by various scientists, for it is the way in which
objectives are understood that determines how they are pursued and the
relevance of accomplished scientific work to social problems. It is
here, to speak only of the domain of biology, that matters of how the
biological world is to be perceived and understood become so important.
For one view of the biological world will lead to very different inter-
pretation and pursuit of even carefully defined objectives than will
another, and these will lead to different relevances of the work to
'scientific and social problems. It is for this reason that we considered
the nature of natural ecological systems, as seems apparent to us, in
Section II of this article. And there we went so far as to articulate an
abstract conceptual framework for biology and to interpret it at ecological
levels of organization. We will not here argue whatever adequacy this
view of the biological world and ecological systems may or may not have.
Here, it is only important that the basis for judgments we make as to the
meaning of defined objectives and their relevance to solutions of problems .
108
-------
in environmental toxicology has been made explicit, so that those concerned
can evaluate the rationality of those judgments based on our presuppositions.
We judge the levels of organization actually represented by laboratory
ecological systems and the capacities of the representational models to
underlie possible interpretation of defined objectives and thus to under-
lie the relevance of these objectives and work done in their pursuit to
problems in environmental toxicology.
Let us first consider what levels of biological organization the
models being called laboratory ecosystems may represent. If, say six or
seven species representing the major trophic kinds—plants, herbivores,.
omnivores, carnivores, and decomposers—are indeed so adapted to one
another and to the system as a whole that this whole has its own system
capacities, what level of natural organization may we suppose this labora-
tory ecosystem represents? Some a priori, view of the nature of natural
ecosystems is necessary to respond to this question. If, for purposes of
argument, one accepts our view of the organization of natural ecosystems,
then what some of us have been calling laboratory ecosystems represent—in
terms of dimensions and also in terms of capacities—at most very low
levels of ecosystem organization. This must have something to do with the
capacities, the performances, and what we can conclude about intoxication
of natural ecosystems, on the basis of laboratory ecosystem studies. This
is so no matter what our particular objectives may be—evaluation of rela-
tive toxicity, evaluation of ecological magnification and biological
degradation of toxicants, or determination of toxic effects on individuals,
populations, and communities. The term ecosystem, as employed in
ecology, has generally meant a natural biological community together with
its physical and chemical environment. No one supposes laboratory ecosystems
have either the dimension or the capacities of natural ecosystems. But we
must be careful not to attribute directly to natural ecosystems the
capacities and performances of laboratory ecosystems, use of the term
laboratory ecosystem perhaps suggesting more relevance than we have any
reason to assume.
Now, of course, laboratory models, just as theories, are representa-
tional idealizations intended to aid our search for understanding of
nature. Good models, just as good theories, must not simply reflect
nature in all the dimensions of her complexity but must somehow represent
that which is most important in determining the behavior of natural systems.
It is not the reduction of dimensions of laboratory ecosystems that may
vitiate their use in environmental toxicology, but whether dimensions have
been reduced in such a way as to leave the models representing nothing at
all of importance in natural ecosystems. It is the capacity of a system
that determines the environments in which it could possibly persist and
what its performances can possibly be in any particular environment. High
level systems have their own level-specific capacities determined in part
by the capacities of their subsystems (Fig. 4)„ Move than anything elset
our laboratory ecosystems must be designed so as to possess capacities
representative of capacities of low-level subsystems of natural eco-
systems; in no way can they represent the capacities of natural ecosystems
as wholes. But if we can demonstrate that even lew level systems have
capacities for ecological magnification and biological degradation and that
109
-------
capacities for persistence or production are decreased by toxic actions, it
is conceptually and theoretically sound to conclude that' the capacities
of natural ecosystems will entail such capacities and alterations of capaci-
ties of their lower-level subsystems. The distinction between capacity
and performance of systems here becomes of the greatest importance for
although there is general interest in the capacities of laboratory multi-
species systems, there is little general interest in their particular per-
formances. This is because not only the capacities but the environments of
laboratory multispecies systems are different from those in nature, and there
is no reason to suppose that magnitudes of particular performances determined
in the laboratory correctly reflect magnitudes of the performances in nature.
For these reasons, we have placed great emphasis on the capacities of indi-
vidual organisms, populations, and multispecies systems in the interpretation
of objectives for laboratory ecosystem research in environmental toxicology
(Fig. 31). Understood in this way, we believe these objectives to be quite
relevant to problems in environmental toxicology.
But application of results of laboratory ecological system studies"to
the solution of problems in environmental toxicology cannot be direct or
simple, for the relationships between knowledge obtained in the laboratory
and the nature of the natural systems in which we perceive problems are
indirect and complex* For this reason, in the previous section we empha-
sized our view that the immediate explanatory objective of laboratory
ecological system studies should be development of sets of empirical generali-
zations. Particular generalizations, based on the results of laboratory
ecological system studies of particular designs, should state as generally
as possible the behavior and effects of representative toxicants in such
laboratory systems. But the generalizations must in some way specify the
nature of the laboratory model yielding the results from which the
generalizations were induced. This must be done because of the
uniqueness of a particular laboratory ecological system. To deal with the
problem of such uniqueness, and thus with the problem of limited generality
of particular generalizations, we must study many different kinds of labora-
tory ecological systems. We must, of course, study laboratory systems
having different levels of biological organization: individual organism,
population, and multispecies systems. But, even at each general level of
organization, different species, relations, and environmental conditions must
be investigated. Perhaps some few generalizations will pertain to all
such systems, and these will be extremely valuable. But many important
generalizations may not apply to all the systems studied and some generalizations
superficially are likely to appear to be contradictory. Such an appear-
ance-may be owing only to inadequacies of our conceptual and theoretical
framework. Generalizations relating directly to one level of biological
organization will not generally be adequate for direct application to other
levels of organization. Nevertheless, because of the hierarchical
organization of multispecies systems into population and individual level
subsystems, we may be able to see ways of -ordering empirical generaliza-
tions pertaining to the different levels of organization so as to represent
multispecies systems and behavior and effects of toxicants more adequately.
This may best be achieved with scientific deductive systems. But adequate
scientific deductive systems are extremely difficult to create. And some
110
-------
more or less rational and useful way of ordering empirical generalizations,
short of a formal deductive system, may be attainable.
Of course, if we do achieve a theory making possible the deduction of
sets of empirical generalizations, we have made an important advance in
science and its application to solution of problems in environmental toxi-
cology. Such a theory makes possible the explanation and logical unification
of the different generalizations, even ones that might otherwise appear
contradictory. Then we can see the relationships among the generalizations
and the boundaries of explanatory applicability of any one of them. And the
entire deductive explanatory system covers a much broader domain of nature and
environmental problems there occurring. Moreover, especially if the theory
is encompassed by an explicated conceptual framework, we can more clearly
see the assumptions we make when we apply any part of the theory, including
the deduced empirical generalizations, to solution of problems in environ-
mental toxicology.
Finally, it is our strong feeling that we ought not to expect from bio-
logicaL theories, models, and empirical generalizations much in the way of
quantitative prediction of the behavior and effects of toxicants in natural
ecologrcal systems. Certainly our laboratory ecological systems must
generally leave out too much for this to be possible. And empirical gen-
eralizations derived from study of such systems can add nothing more.
Theories do add something more. But the unique capacities and various
environments of organismic systems lead to such vast arrays of kinds and
levels of performances that we are not likely to soon have theoretical
or other ways for achieving precise quantitative prediction. Rather, we
think, it is the kinds of relationships and the forms of relationships
among biological identities and the sorts of changes in these result-
ing from toxicants with which we can hope to deal by theoretical and empiri-
cal means. It is this sort of "qualitative," rather than strictly
quantitative, application of the results of laboratory ecological system
studies that we can reasonably expect to be successful.
Ill
-------
REFERENCES
1. Andrewartha, H. G., and L. C. Birch. The Distribution and Abundance of
Animals. The University of Chicago Press, Chicago. 1954.
2. Bertalanffy, L. von. General System Theory: Foundations, Development,
Applications. George Braziller, New York. 1968.
3. Bisson, P. A., and G. E. Davis. Production of Juvenile Chinook Salmon,
Onoorhynohus tshatfytscha3 in a heated model stream. Fish. Bull. 74
(4):763:774. 1976.
4. Booty, W. M., and C. E. Warren. A General Theory of Productivity and
Resource Utilization, Manuscript.
5. Braithwaite, R. B. Scientific Explanation: A Study of the Function of
Theory, Probability, and Law in Science. Cambridge University Press,
London. 1953,
6. Braithwaite, R. B. Models in the Empirical Sciences, in; Nagel, E.
et al. (eds.). Proceedings of the Congress International Union for
Logic, Method, and Philosophy of Science. 1960.
7. Carson, R. Silent Spring. Houghton Mifflin Company, Boston. 1962.
8. Clements, F. E. Plant Succession. Carnegie Institute of Washington
Publication 242, Washington. 1916.
9. Cole, L. K., J. R. Sanborn, and R. L. Metcalf. Inhibition of Corn Growth
by Aldrin and the Insecticide's Fate in the Soil, Air, Crop, and Wild-
life of a Terrestrial Model Ecosystem. Environmental Entomology
5:583-589. 1976.
10. Fry, F. E. J. Effects of the Environment on Animal Activity. University
of Toronto Studies Biological Series 55, Ontario Fisheries Research
Laboratory Publication 68. 1947.
11. Gleason, H. A. The Structure and Development of the Plant Association.
Bulletin of Torrey Botanical Club 44:413-481. 1917.
12. Gleason, H. A. The Individualistic Concept of the Plant Association.
Bulletin of the Torrey Botanical Club 53:7-26. 1926.
112
-------
13. Gillett, J. W. and Jay D. Gile. Progress and Status Report on Terres-
trial In-house System. Proceedings of "Substitute Chemicals Program -
The First Year of Progress." 1975.
14. Gutting, G. Conceptual Structures and Scientific Change. Studies in
History and Philosophy of Science 4:209-230. 1973.
15. Hanson, N. R. Patterns of Discovery. Cambridge University Press,
Cambridge. 1958.
16. Isensee, A. R. and G. E. Jones. Distribution of 2,3,7,8-tetrachloro-
dibenzo-p-dioxin (TCDD) in Aquatic Model Ecosystem. Environmental
Science and Technology 9:668-672. 1975.
17. Isensee, A. R., P. C. Kearney, E. A. Woolson, G. E. Jones, and V. P.
Williams. Distribution of Alkyl Arsenicals in Model Ecosystem.
Environmental Science and Technology 7:841-845. 1973.
18. Kanazawa, J., A. R. Isensee, and P. C. Kearney. Distribution of Carbaryl
and 3,5-xylyl Methylcarbamate in an Aquatic Model Ecosystem. Journal
Agricultural Food Chemistry 23:760-763. 1975.
19. Kuhn, T. The Structure of Scientific Revolutions. University of Chicago
Press, Chicago. 1962.
20. Lewontin, R. C. The Genetic Basis of Evolutionary Change. Columbia
University Press, New York. 1974.
21. Liss, W. J. The Dynamics-of Exploited Fish Populations Exposed to Diel-
drin. MS Thesis. Oregon State University. 1974. 73 pp.
22. Liss, W. J., and C. E. Warren. On the Possible Form of Future Fisheries
Theory. Manuscript.
23. Lloyd, R. The Toxicity of Zinc Sulfate to Rainbow Trout. Annals of
Applied Biology 49:535-538. 1960.
24. Lu, P-Y., R. L. Metcalf, A. S. Hirwe, and J. W. Williams. Evaluation
of Environmental Distribution and Fate of Hexachlorocyclopentadiene,
Chlordene, Heptachlor, and Heptachlor Epoxide in a Laboratory Model
Ecosystem. Journal Agricultural Food Chemistry 23(5);967-973. 1975.
25. Maguire, B. Jr. Niche Response Structure and the Analytical Potentials
of its Relationship to the Habitat. American Naturalist 107:213-246.
1973.
26. Matsumura, F., and H. J. Benezet. Studies on the Bioaccumulation and
| Microbial Degradation of 2,3,7,8-tetrachlorodibenzo-p-dioxin.
i) Environmental Health Perspectives 5:253-258. 1973.
113
-------
27. Metcalf, R. L. Screening Compounds for Early Warnings about Environ.,
mental Pollution. Proc. Univ. Mo. Annual Conf. Trace Subst. Environ.
Health 8:213-217. 1974a.
28. Metcalf, R. L. A Laboratory Model Ecosystem to Evaluate Compounds Pro-
ducing Biological Magnification. Pages 17-38. In W. J. Hayes, Jr. (Ed.),
Essays in Toxicology, Vol. 5. Academic Press, New York, xii + 189 pp.
1974b.
29. Metcalf, R. L. A Laboratory Model Ecosystem for Evaluating the Chemical
and Biological Behavior of Radio-labelled Micro-pollutants. Pages 49-63.
In E. R. A. Beck (Ed.) Comparative Studies of Food and Environmental Con-
tamination. Proceedings International Atomic Energy Agency Series.
1974c.
30. Metcalf, R. L., I. P. Kapoor, P-Y. Lu, C. K. Schuth, and P. Sherman.
Ecosystem Studies of the Environmental Fate of 6 Organo Chlorine Pes-
ticides. Environmental Health Perspectives 4:35-44. 1973,
31. Metcalf, R. L. and J. R. Sanborn. Pesticides and Environmental Quality
in Illinois. Illinois Natural History Survey Bulletin 31:381-436. 1975.
32. Metcalf, R. L., G. K. Sangha and I. P. Kapoor. Model Ecosystem for the
Evaluation of Pesticide Biodegradability and Ecological Magnification.
Environmental Science and Technology 5:709-713. 1971.
33. Morgan, J. P. Effects of Aroclor 1242R (a Polychlorinated Biphynl) and
DDT on Cultures of an Alga, Protozoan, Daphnid, Ostracod, and Guppy.
Bulletin Environmental Contamination and Toxicology 8:129-137. 1972.
34. Mount, D. I., and C. E. Stephan. A Method for Establishing Acceptable
Toxicant Limits for Fish — Malathion and the Butoxyethanol Ester of
2,4-D. Transactions of the American Fisheries Society 96:185-193. 1968.
35. Nagel, E. The Structure of Science. Harcourt, Brace, and World, New
York. 1961.
36. Popper, K. The Logic of Scientific Discovery. Basic Books, New York.
1959.
37. Ramensky, L. G. Die Grundgesetzmassigkeiten im Aufbau der Vegetations-
decke. Wjenstn. opytn. djela Woronesch. (Bot. Centralbl., N. F.
7:453-455). 37 pp. 1924. '
38. Roelofs, T. D. Effects of Dieldrin on the Intrinsic Rate of Increase of
the guppy, Poeoilia vetiaulata Peters. Ph.D. Thesis Oregon State Univ-
ersity, Corvallis. 1971.
39. Sanborn, J. R., and C-C. Yu. The Fate of Dieldrin in a Mo'del Eco-
system. Bulletin Environmental Contamination and Toxicology
10:340-346. 1973.
114
-------
40. Seim, W. K., J. Lichatowich, R. Ellis, and G. Davis. Effects of Kraft
Mill Effluents on Juvenile Salmon Production in Laboratory Streams.
Water Research 11(2) :189-196. 1977.
41. Shumway, D. L., C. E. Warren, and P. Doudoroff. Influence of Oxygen
Concentration and Water Movement on the Growth of Steelhead Trout and
Coho Salmon Embryos. Transactions of the American Fisheries Society
93:342-356. 1964.
42. Silliman, R. P. Interaction of Food Level and Exploitation in Exper-
imental Fish Populations. Fisheries Bulletin of Fish and Wildlife
Service 66:425-439. 1968.
43, Taub, F. B. Biological Models of Freshwater Communities.
EPA-660/3-73-008. 1973.
44. Thatcher, T. 0. Some Effects of Dissolved Oxygen Concentration on
Feeding, Growth and Bioenergetics of Juvenile Coho Salmon. Ph.D.
Thesis, Oregon State University, Corvallis. 1975.
45. Toulmin, S. Philosophy of Science. Harper and Row, New York. 1960.
46, Warren, C. E. Biology and Water Pollution Control. W. B. Saunders,
Philadelphia. 1971.
47. Warren, C. E., W. K. Seim, R. 0. Blosser, A. L. Caron, and E. L. Owens.
Effect of Kraft Effluent on the Growth and Production of Salmonid Fish.
Tappi 57(2):127-132. 1974.
48. Warren, C. E., M. W. Allen, and J. H. Haefner. Conceptual Frameworks
for Biology: A Tentative Proposal. Manuscript.
49. Warren, C. E., and G. E. Davis. Laboratory Studies on the Feeding,
Bioenergetics, and Growth of Fishes. In S. D. Gerking (ed.), The Biol-
ogical Basis of Fish Production. Blackwell Scientific Publications,
Oxford. 1967.
50. Warren, C. E., and G. E. Davis. Laboratory Stream Research: Objectives,
Possibilities, and Constraints. Annual Review of Ecology and Systematics
2:111-144. 1971.
51. Wurtsbaugh, W. A. Effects of Temperature, Ration, and Size on the
Growth .of Juvenile Steelhead Trout, Salmo gairdneri. MS Thesis, Oregon
State University, Corvallis. 1973.
115
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/3-77-022
3. RECIPIENT'S ACCESSIOf*NO.
4. TITLE AND SUBTITLE
Design a.nd Evaluation of Laboratory Ecological System
Studies
5. REPORT DATE
December 1977 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Charles E. Warren and William J. Liss
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Fisheries and Wildlife
Oak Creek Laboratory of Biology
Oregon State University
Corvallis, Oregon 97331
10. PROGRAM ELEMENT NO.
1BA608
11. CONTRACT/GRANT NO.
Grant No. .
R802286
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
14. SPONSORING AGENCY CODE
EPA-600/03
16. SUPPLEMENTARY NOTES
16. ABSTRACT
Design and evaluation of laboratory ecological system studies are considered in
relation to problems and objectives in environmental toxicology. Ecological systems
are defined to be organismic systems together with their level-specific, co-extensive
environmental systems and to occur at individual, population, and multispecies levels
of biological organization. So that the basis for judgments on the relevance and
adequacy of laboratory ecological system studies for solution of problems in environ-
mental toxicology will be clear, a conceptual framework defining with abstract
generalisations the nature of biological systems is presented and employed. And a
graphical calculus is used to deduce isocline systems and dynamic as well as steady-
state behaviors of multispecies systems, so as to illustrate the importance of
empirical evaluation of the capacities, not simply the performances, of laboratory
ecological systems. Within the context of apparent toxicological problems and this
conceptual framework, the relevance and adequacy of laboratory ecological system
studies on toxicant effects and behaviors are evaluated.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
models
environment- simulation
closed ecological systems
aquatic biology
communities
populations
pollution
toxicology
laboratory
microcosms
06 C,F,T
14 B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
i I £. '
20. SECURITY CLASS (Thispage}
UNCLASSIFIED
22. PRICE
EPA Form 2220-1! (9-73)
ft U.S. HOTKNMEKT WHNTOIG OFFICt 1978- 7 S7-140/6672
------- |