A -OOMIH)
JUN
no?niY or
.m PROTECTS won
.ARY RESIOH X
1200 SIXTH AVEKOE
, WASH. Wit
THE USEFULNESS OF BIOLOGICAL
COMMUNITY INDICES FOR ENVIRONMENTAL
STANDARDS, CRITERIA, AND
ENFORCEMENT - PART I
EPA-LAG-D4-F461
NATIONAL ECOLOGICAL RESEARCH LABORATORY
An Associate Laboratory of
National Environmental Research Center—Corvallis
-------
THE USEFULNESS OF BIOLOGICAL COMMUNITY INDICES
FOR ENVIRONMENTAL STANDARDS, CRITERIA, AND
ENFORCEMENT - PART I
QUESTIONAIRRE SURVEY OF RESEARCHERS AND
LITERATURE REVIEW
NATIONAL ECOLOGICAL RESEARCH LABORATORY
U.S. ENVIRONMENTAL PROTECTION AGENCY
CORVALLIS, OREGON 97330
JULY 1974
EPA-IAG-D4-F461
.JS3.HS!HW»™ 10 MATERIALS
RXDDDOE43Sfl
-------
ABSTRACT
An Investigation was conducted by written Inquiry and a search of relevant
literature to determine common criteria for the acceptability and use of
biological community parameters, especially indices of species diversity,
to enforce pollution regulations in the terrestrial, freshwater, and marine
habitats. The investigation revealed that while most respondents from the
queried industrial sector, regulatory agencies, and researchers were of the
opinion that biological communities were extremely important, there was
little agreement on means of measurement. Attempts at application of
specific theoretical distributions describing species diversity have given
differing results. It is recommended that environmental criteria based on
species diversity not be established at the present time.
Richard Vanderhorst
Peter Wilkinson
ii
-------
CONTENTS
Page
Abstract il
List of Tables iv
Sections
I Conclusions ]
II Recommendations 3
III Introduction 4
IV Written Inquiries 6
V Literature 15
VI References 37
VII Appendices 44
iii
-------
TABLES
No. Page
1 Characterization of Typical Respondent by Major Group 7
2 Characterization of Typical Respondent by Habitat of
Principal Involvement 9
3 Characterization of Typical Respondent by Number of
Years' Experience in Environmental Assessment 10
4 Average Number Indicators Selected by Respondents 12
5 Three Hypothetical Communities Having the Same Number 19
of Species (S) and Total Number of Individuals (N)
That Yield the Same Diversity Index, d~ = S-l/logeN
6 Examples of Species Diversity, d_, in Polluted Waters 22
7 Comparison of Mean Annual cf (diversity per individual) 23
Values per Station
8 Species Diversity Values for the Samples Under Observation 27
as Obtained from Different Diversity Indices
9 Correlation Coefficients of Different Pairs of Diversity 28
Indices
IV
-------
SECTION I
CONCLUSIONS
The following conclusions may be drawn with respect to the use of biologi-
cal community parameters, especially species diversity, for applied, or
enforcement criteria:
Although there is a common concern for biological communities among the
three major sectors potentially affected by environmental quality cri-
teria, there is little agreement as to how best to numerically express
change in community structure. This disagreement is not along associa-
tive group lines.
A major obstacle to such agreement is the lack of acceptance of a common
definition for the term "community."
"Species diversity" has also been defined and used ambiguously and such
ambiguity is not resolved. Recent workers have independently expressed
two distinct aspects of species diversity, i.e., species richness and rel-
ative distribution among species.
Numbers of species per unit area/volume most meaningfully describes species
richness.
Computation of any given index may render a result either consistent or
inconsistent with other measures of environmental quality.
If indices of species diversity are to be useful as criteria they must be
considered on a case-by-case basis.
Information derived from presently available indices of species diversity
is not biologically interpretable.
Information derived from presently available indices of species diversity
is not convertible into an economic denominator for derivation of cost versus
benefit.
Presently available indices of species diversity do not convey information
regarding the esthetic value of biological associations.
-------
Habitat lines drawn as terrestrial, freshwater, or marine, do not
relate to the usefulness or lack of usefulness of a given index.
The Overwhelming evidence from the survey of opinion by written in-
quiry and examination of published information in the present under-
taking suggests that we are not ready at this time to asstgn specific
magnitudes for specific theoretical distributions as environmental
criteria.
-------
SECTION II
RECOMMENDATIONS
It 1s recommended that presently available indices of species diversity
not be used as regulatory criteria.
It is recommended that the information theory formula and equitability
49
component as tabulated by Lloyd and Ghelardi. be computed in on-going
U. S. Environmental Protection Agency investigations where good supportive
data of other types are collected. The value of these computations would
lie in the validation of the indices rather than drawing inferences about
environmental quality from index magnitudes.
The U. S. Environmental Protection Agency should adopt a working defini-
tion of the biological community. The definition should encompass the
total biotic assemblage within physically defined boundaries.
The U. S. Environmental Protection Agency should encourage comparative
investigations of the biological community.
-------
SECTION III
INTRODUCTION
In the past decade a plethora of literature has grown up concerning
the theoretical and practical value of biological community parameters,
especially indices of species diversity, for assessment of environmental
well being. Proponents of the use of such indices have emphasized the
demonstrated ability to produce a numerical, dimensionless index which
retains historical information concerning a biological assemblage and
the effects of a pollutant or environmental insult on that assemblage
as opposed to physical and chemical indices which reflect only on the
instantaneous condition of the environment surrounding the assemblage.
Efforts by many investigators have centered on simplifying the calcula-
tion of such indices to make possible their estimation and interpretation
by persons without a technical biological background. Opponents of the
use of species diversity indices are generally not opposed to the concept
of species diversity or to the biological community concept, but do ex-
press concern that simple number indices of any kind are apt to be mis-
leading in interpretation. Further, there is agreement that confusion
exists with respect to the definitions of "communities" and "diversity."
The U. S. Environmental Protection Agency has requested an assessment of
the use and acceptance of biological community parameters, especially in-
dices of species diversity, for evaluating environmental change induced
by pollution stress and the development of criteria for the use of such
indices for enforcement purposes. Battelle Pacific Northwest Laboratories
conducted the here-reported preliminary investigation toward the following
objectives:
Determine by written inquiry the acceptability by researchers, by
industry, and by government agencies, the use of species diversity
indices, alone, or in aggregate, for use as enforcement tools.
Determine by a search of the relevant literature background material
on the approaches and uses made of community parameters for environ-
mental assessment.
-------
Assimilate criteria for the use of community parameters, especially
species diversity indices, in the terrestrial, freshwater, and
marine environments.
Recommend future courses of action for the development of community
parameters for environmental assessment and enforcement.
-------
SECTION IV
WRITTEN INQUIRIES
In an effort to ascertain current understanding and level of acceptance of
biological conmunity analysis as a measure of environmental change, an in-
quiry was circulated among persons and groups directly involved with environ-
mental study. To get a reasonable cross section of opinion, we felt it
would be appropriate to divide involved groups into three broad classifica-
tions: the industrial sector; regulatory agencies; and research groups
representing the relevant technical expertise.
An inquiry format was developed which, while somewhat subjective.was felt
to be a practical approach to summarizing data obtained from a diverse
collection of groups and individuals having widely differing experience
with biological communities. We had hoped the format (see_ Appendix A)
was lucid enough to elicit responses from people ranging from those with
limited knowledge and understanding of the subject through those who are
involved in theory development. That we were able to achieve some measure
of success in the above outlined objective is indicated by the large number
of thoughtful responses received.
Mailing lists were obtained from indices of ongoing research in relevant
fields, of regulatory agencies, and of industries likely to be involved in
environmental studies. Copies of the inquiry were mailed to individuals
representing each of the three sectors. Three thousand two hundred and
twenty-six inquiries were mailed to groups in the United States, Puerto
Rico, and Canada.
RESULTS
Ten percent (334) of the respondents completed the inquiries and returned
them by the 1 June deadline.
In order to Identify a typical respondent we have classified them in three
ways. Table 1 lists the associative group, i.e., industrial, regulatory, or
research. The column information on the table results from a majority re-
sponse to the "type of involvement" question. For example, the principal
-------
Table 1. CHARACTERIZATION OF TYPICAL RESPONDENT BY MAJOK GROUP
Type of involvement
Principal habitat
Level of participation
Understanding
Position held
Years experience
Section I. Identification
Associative group
Industry
Freshwater
Full time
Moderate
Other
10 or more
Regulatory agency
Freshwater
Full time
Comprehensive
Biologist-ecologist
10 or more
Research
Terrestrial
Frequent
Comprehensive
Biologist-ecologist
10 or more
Section II. Type of involvement
Criteria used
Frequency of environmental
assessment
Percent using listed
index (one or more)
Indicator of change
Number of parameters
Recommend for limited
training?
Other indices used?
Rating of community analysis
Field survey
Occasionally
66%
Section III,
Communities
More than one
No
Yes
Good
Field survey
Occasionally
84%
Opinion
Communities
More than one
No
Yes
Good
Field survey
Occasionally
90%
Communities
More than one
No
Yes
Good
Characterization derived from percentage response to question of interest. For details see Appendix B.
-------
habitat of most industrial and regulatory respondents was freshwater
while the researchers were predominantly terrestrial in specialty.
Table 2 characterizes a typical respondent in terms of habitat of prin-
cipal involvement. A substantial portion of respondents were undecided
as to the habitat of principal involvement (see_ Appendix C for detail).
Thus a column for undecided individuals is included. By way of example,
from Table 2, terrestrial and freshwater workers reported their frequency
of involvement in environmental assessment as occasional; marine workers
reported their involvement as frequent; and, the majority of undecided re-
spondents reported they were never involved in environmental assessment.
A third classification for respondent characterization, based on years of
experience in environmental assessment, is presented on Table 3. Classes
were established as zero to two, two to five, five to ten, and greater than
ten years. As for characterization by habitat a substantial portion of
the respondents were undecided and an appropriate category is included.
In preparation of the inquiry we listed several common indices of species
diversity. That part of the inquiry was designed first, to elicit re-
sponse about the use of those particular indices, and second, to stimulate
suggestion of other indicators. There were a few respondents who objected
to the approach, but on the whole the response was excellent. Data on the
use of the indices listed on our inquiry are presented on Tables 1 to 3
for each of the specific classifications. A selection of additional sug-
gestions by respondents follows:
Method No. responses
Analysis of variance 1
Organlsms/sq.ft. 1
Community metabolism 1
Sequential comparison index 4
Growth rates 1
Serological analysis 1
Long-term study plots 1
Shannon-Wiener H1 8
Successional patterns 1
-------
Table 2. CHARACTERIZATION OF TYPICAL RESPONDENT BY HABITAT OF PRINCIPAL INVOLVEMENT
Type of involvement
Criteria used
Frequency of environmental
Percent using listed index
(one or more)
Indicator of change
Number of parameters
Recommend for limited training?
Other indices used?
Rating of community analysis
Section II. Type of involvement
Principal
Terrestrial Freshwater
Field survey Field survey
Occasionally Occasionally
94% 89%
Section III. Opinion
Community Community
More than 1 More than 1
No No
Yes Yes
Good Good
involvement
Marine
Field survey
Frequently
93%
Combination
More than 1
No
Yes
Good
Undecided
Field survey
Never
58%
Combination
More than 1
No
Yes
Inconclusive
For details see Appendix C.
-------
Table 3. CHARACTERIZATION OF TYPICAL RESPONDENT BY NUMBER
OF YEARS' EXPERIENCE IN ENVIRONMENTAL ASSESSMENT*
Section II. Type of involvement
Years of experience
Type of involvement
Criteria used
Frequency of environmental
assessment
Percent using listed index
(one or more)
Indicator of change
Number of parameters
Recommend for limited
training?
Other indices used?
Rating of community analysis
0-2
Field survey
Occasionally
65%
Section
Community
More than 1
No
No
Inconclusive
2-5
Field survey
Occasionally
54%
III. Opinion
Community
More than 1
No
Yes
Good
5 - 10
Field survey
Occasionally
QOV
OOyb
Community
More than 1
No
Yes
Good
10+
Field survey
Occasionally
92%
Combination
More than 1
No
Yes
Good
Undecided
Inconclusive
Occasionally
33%
Inconclusive
More than 1
No
Inconclusive
Good
Responses falling on a division were elevated to the next higher rank. For details see Appendix D.
-------
Method No. responses
Direct observation (subjective) 2
Trap-day success 1
Mammalian density 1
Coefficient of similarity 2N/A+B 2
Floristic inventories 2
Cluster analysis 2
Recurrent group analysis 1
Pielou evenness (J) 1
Dissimilarity index 1
Productivity tests 3
Empirical biotic index 1
Sturber - pollution tolerance 1
Trophic index 2
Peterson index 2
Indicator organisms 5
Margalef - chlorophyll ratio 1
Length weight ratio 1
Population structure 1
MacArthur - multiple M 1
Manipulated sample populations 1
Index of faunal affinity 1
Equitability - Lloyd and Ghelardi 1
Ordination analysis 1
Stream drift standing crop 1
Least squares method 1
Paleoecological indicators 1
Rest species increase 1
Community resilence measures 1
Principal component analysis 1
Variance pattern 1
Relative abundance 1
11
-------
In response to choice of indicators of environmental change, the average
number of indicators used by each group are presented in Table 4. For all
classifications, a general response is that two or more indicators should
be used.
Table 4. AVERAGE NUMBER INDICATORS SELECTED BY RESPONDENTS
Number
indicators
i Number
indicators
Number
indicators
Industry
2.3
Terrestrial
2.7
0-2 2-5
3.3 2.0
Regulatory
2.3
Fresh
2.5
5-10
2.5
agency
Marine
2.7
10+
2.7
Research
2.5
Undecided
2.5
Undecided
4.0
DISCUSSION
Examination of response data reveals four main conclusions:
The majority of respondents feel they understand what a biological com-
munity is.
Most respondents feel community response is the best indicator of en-
vironmental change.
There is no general agreement on how to measure community response or what
the significance is of the results of indices used.
A substantial portion of respondents expressed doubt that any current
index or other numerical expression of community response can replace
judgement on the part of the investigator.
With reference to level of understanding of commum'ty response the conclusion
is derived from a simple tally of responses to direct question. Admittedly
there is likely some bias in the result because of a natural desire of the
respondents to indicate broad knowledge.
12
-------
In answering Question 1, Section III, most respondents Indicated "the
community" is the best indicator of environmental change. The majority
of respondents chose more than one indicator.
Many of the respondents wrote marginal remarks, especially on Question 1,
Section III: "An indicator of environmental change should be mainly:"
Typical responses were:
- All of the above
- It depends on the community
- All of the above and then there is doubt if that is enough
- Change in energy flows
- Site dependent
- ...We need integrated studies, not indicators
- Short term changes may be indicated by chemical analysis and long
term by the biological community
- The community is probably the most important, however, chemical
analysis and information at the organism and population level is
also important
- Should consider lethal and sublethal effects on the organism and
population as related to the total community.
- Ecosystem parameters
- Should not be mainly chemical; but should include chemical,
community and other biological evaluations as feasible
- Biochemical analyses of tissues of the organism
- Each situation should be assessed - there is no one answer
- No generalization possible or desirable
- Depends on the community (or environment)
- Total systems response
- Freshwater community, taking into consideration entire watershed
ecosystem
- Community metabolism
- All 3 (sic) above used with extreme care
- Combination - depending on the type of questions needing answers
- Each area is different and'should be treated as a living system
with chemical parameters showing the changes and biological the
13
-------
reasons
- Community is ultimate, but chemistry and the organism are essential
to determine early problems and trends
The above-listed comments are in our judgement typical of the subjective
comments to the question posed. The inescapable conclusion from these
comments is that the respondents feel that ecosystems are far too complex
to generalize about in terms of a single numerical index. The great pre-
ponderance of respondents encouraged the use of community response in en-
vironmental assessment; however, there was little agreement on methods of
measurement. Two concepts were stressed to the point of redundancy. First,
judgement on the part of the investigator is paramount in evaluating com-
munity response. Second, programs must be tailored to individual situations.
14
-------
SECTION V
LITERATURE
The objective of the here-reported review of literature is not to supplant
1 2
the many excellent reviews, e.g., Woodwell and Smith , Connell and Orias ,
Whittaker , MacArthur , Pianka , and Mclntosh , concerning the use of spe-
cies diversity in specific habitats for evaluation of community change.
Rather, we attempt to bring together common problems for the establishment
of inter-habitat criteria for the assessment by regulatory agencies of
potentially deleterious changes brought about by the activities of man.
The biological community, at a point in time, is a reflection of the phy-
sical, chemical, and biological components of its history. Insofar as it
retains a record of events leading to development, the community, by reason
of its integrating, or information retaining ability, forms the basis for
ah environmental monitoring tool far superior to those methods involving
solely the measurement of physical and chemical variables. If one accepts
this opening premise, then there are three conditions that must be ful-
filled if we are to establish environmental quality criteria in terms of
diversity of the biological community. First, we must understand what
events in the community's development contribute to the qualitative struc-
ture of the community. Second, we must find an adequately concise means
of quantitatively expressing the community's qualitative features; and
third, we must assign a reasonably derived acceptable level of community
development for the criterion. Most of the literature reviewed for this
study related to the second of these three conditions.
Even the strongest advocates of using diversity indices as water quality
criteria concur that data are needed on specific communities before the use-
fulness of the index is evident. Thus, Wilhm and Dorris state, "After a
particular diversity expression is accepted and a meaningful agreement is
found between the natural community and a theoretical distribution, a char-
acteristic diversity value can be found to express the structure of each
community." And, later in the same paper they state, "Additional work
t
needs to be done to learn how types and degrees of pollution are expressed
p
in d_." Cairns and Dickson , in presentation of a simple method for biological
15
-------
assessment, state: "Additional work must be done to learn how different
types and degrees of pollution are expressed in terms of DIj."
The term "community" has been defined in a variety of ways. Thus, defi-
nitions for community include those which range from a simple assemblage
of all the biota at a given location to patterns of natural assemblages
having cohesion, and perhaps, culminating in definitions which consider
the community as a superorganism. In short, the concept of biological
community is generally recognized and accepted but not consistently de-
g
fined. Further, consistent with Warren's view, our feeling is that
pollution is fundamentally a social problem. Hence, the biologist's
role becomes that of adequately translating technical descriptions of
environmental change into a socially recognizable form. Recommended levels
for a particular index or description should relate not to desirability or
undesirability but rather to the most accurate description of extant condi-
tions.
The definition of "diversity" is equally clouded in the ecological literature.
Example definitions can be found in Hill10, Pielou11, Wilhm12, and Whittaker3'
13
Hdrlbert states: "Species diversity has been defined in such various and
disparate ways that it now conveys no information other than 'something to
do with community structure.'"
Nevertheless, those charged with the responsibility of establishing criteria
for environmental assessment, or enforcing such criteria, or complying with
such criteria must have a common reference. Definitions aside, there does
seem to be common agreement that something called a community is extremely
important. Arguments relating to the community's status as a level of or-
ganization, the degree of interacting functionality, or lack of it, do not
seem to detract from the need to monitor and understand the effects of man's
activities on biological communities. That communities are important leads
us to the position that the U. S. Environmental Protection Agency should
adopt a working definition of "community." A common problem, and indeed,
14
one exemplified by EPA is the use of the term community in two different
16
-------
14
ways in the same paper. Thus on page 16, EPA , "Diversity indices are an
additional tool for measuring the quality of the environment and the effect
of induced stress on the structure of a community of macroinvertebrates"
(italics mine). And later, on the same page, "These confounding factors can
be reduced by comparing diversity in similar habitats and by exposing arti-
ficial substrate samplers long enough for a relatively stable, climax com-
munity to develop." (italics mine). The former, limited to a few populations,
i.e., the macroinvertebrates, would be hard to visualize as a functional
entity. The latter, on the other hand, could readily be visualized in such
a manner.
The pragmatic question of what one is measuring when evaluating results
from "community" studies is related to the problem of community definition.
Thus, thoughtful individuals, e.g., Hill , Eberhardt (personal communication),
have aptly pointed out that analyses are applied to collections, or samples,
and not to natural assemblages. Typically in the literature inferences are
drawn about natural assemblages when in fact data are totally lacking to
place the samples or collections in,to perspective with respect to the nat-
ural assemblages. Those collections, by inference, are loosely referred to
as "communities." We suggest that an interim working definition for com-
munity include reference to the total natural assemblage, and for practical
purposes, that physically defined limits be placed as boundaries of the
defined assemblages. Alternatively, we would propose that the term "com-
munity" be dropped entirely from the language of regulations and guidelines.
Indices of species diversity are said to relate to the structure of commu-
nities by almost all investigators(Fisher, Corbet^and Williams , MacArthur
and Wilson , Margalef , Patten , Pianka , Pielou , Sanders , Shannon
20 21 3
and Weaver , Simpson , Whittaker , and many others). The biological
meaning of the indices initially was closely allied to the relationship be-
tween diversity and stability, i.e., the more diverse community is stable;
and hence, the community is better adapted. In considering such a relation-
19
ship, however, one is soon enmeshed in evolutionary time. Thus, from Sanders '
"It requires appreciable time to evolve a highly diverse fauna, and the time
17
-------
component of our stability-time hypothesis is perhaps best illustrated with
lakes. Most lakes are of a relatively transitory nature, or of recent geo-
logic origin. It has been 10,000 years or less since the last glaciation,
and the aquatic fauna from such recently glaciated regions shows limited
diversification" (italics mine).
Sanders goes on to say that ancient lakes, 30 million years old, are charac-
terized by a highly diverse fauna. The seemingly innumerable biological
trials and errors that must have prevailed in the development of a "diverse"
fauna are difficult for our imagination to fathom. The "jump" to instan-
taneous evaluations based on an index, for determination of the effects of
a man-related accident or even, in fact, what we commonly call chronic dis-
turbances (several years in duration) does not seem justifiable in terms of
a stability related theory. If one accepts the foregoing line of thought
then arguments relating to the definition of "diversity" should not muster
support from a stability-theoretical base but rather should be rooted in
the empirical usefulness of the measure employed. That position is con-
sistent with Hill's , discussion of diversity in terms of samples with
a total separation from thermal dynamic feedback and information theories.
22
Eberhardt's position was that an adequate number of theoretical distribu-
tions were available but that data on the representativeness of sampling
for natural assemblages had received little attention. The Shannon-Wiener
function or modification therefrom has received by far the most use both
for theoretical and applied purposes. In our view, the index has the dis-
tinct disadvantage of beijjg tied inappropriately to: (1) community diversity-
stability theory and (2) communication engineering theory. We have already
made some statements with respect to the former. With respect to the latter,
23 '
a comment from Gilbert seems appropriate, "As an engineering subject, in-
formation theory has flourished for 18 years because of the promise it gave
of Improved communication systems. The results are still almost exclusively
on paper." Shortly we will present some of the empirical data which have
been generated on information theory indices in applied situations but it
seems most appropriate to review the salient features in the development of
species diversity indices.
18
-------
In simplest form a diversity index is the ratio of number of species
to number of specimens, expressed:
S/N
where, S = number of species and N = the number of specimens. Large
variations in numbers of specimens due to clumping, season, and sample
size led quickly to the use of a damping function in the denominator so
that the index becomes:
d = S-l/logeN or d = S-
where,• d = species richness, and S and N are defined as above.
A table from Wilhm and Dorris aptly illustrates the problem of i
values for such an index:
Table 5. THREE HYPOTHETICAL COMMUNITIES HAVING THE SAME
NUMBER OF SPECIES (S) AND TOTAL NUMBER OF INDIVIDUALS (N)
THAT YIELD THE SAME DIVERSITY INDEX, d = S-l/logeN*
Individuals in species i (n.. )
Community
A
B
C
nl
20
40
96
n« n,
& 3
20 20
30 14
1 1
7
n4
20
10
1
n5
20
5
1
Total
individuals
N
100
100
100
Total
species
S
5
5
5
*After Wimm and Dorris
Wilhm and Dorris correctly point out that the hypothetical communities are
very different in structure but that equivalent values are obtained for N,
S, and d". Thus the d = species richness, from Margalef , becomes recognized
as a component of diversity and the need apparently is to devise an index
which encompasses other components; most pressing, perhaps, is the component
of evenness of distribution of the specimens among the species.
19
-------
The index of species richness has been shown by numerous authors to in-
crease with increasing sample size and has been criticized for that fea-
Q
ture. As Warren points out the relative abundance of species (species
richness) certainly influences our idea of diversity and should be in-
cluded in an expression of diversity. However, we take the position
that relative abundance expressed in terms of species per individuals
collected can be so dramatically influenced by conditions of the moment
(e.g., clumped sample; large hatch or spawn) as to be uninterpretable as
an index to community condition. Numbers of species per unit of area/
volume would seem to convey the idea of "richness" much more clearly.
25
Margalef is credited with being the first to propose that species
diversity indices be based on information theory. The most commonly
used expressions are the Shannon-Wiener function:
S
H = - Vr. Iog2 TT^ 1n the form.
S
h = - z n./N log, n./N,
1=1 1 * 1
where, H-= entropy, S = number of species;
26
and Brillouln's index:
S
H = (l/N)(log N! - z log Nj).
where the terms are as defined above.
If the ni are large the two formulations give comparable results but in
most cases the n^'s are not large. The information theory formulae are
used to compute the uncertainty concerning the species. The degree of
uncertainty is greater when the diversity is greater. Commonly authors
refer to bits of Information per unit. Further, a propounded advantage
is that the index may be used on'continuous variables such as biomass,
size, or chemical content and is not limited to numbers of individuals
or numbers of species.
20
-------
FRESHWATER
For the investigation of freshwater streams the information theory
type of index has been used more frequently than others. Some authors
have reported high success in obtaining index values in terms of quality
of the receiving stream water, and have, in fact, recommended that water
quality criteria be drawn on specific indices. Other workers have attempted
to apply recommended indices to stream situations with differing results.
7 27
Wilhm and Dorris ' present data on benthie macroinvertebrates of Skeleton
Creek , Oklahoma. Their data show a trend from d" ^ 0.75 at six miles below
where municipal and industrial wastes enter the creek to a cF ^ 3.5 some 61
miles below the outfall. Statistical tests indicated that stations from six
to 32 miles downstream from the outfall were not significantly different but
that stations 43 and 61 miles downstream were significantly different from
the upper stations in mean annual diversity. In the fall the three upper
stations (6, 12, and 16 miles downstream) were significantly different from
stations at 27 an'd 32 miles downstream and in spring stations at 6, 12, 16,
27, and 32 miles downstream were not significantly different from each other
but were significantly different from a station 61 miles downstream. Wilhm
and Dorris present data from other studies in support of the usefulness of
diversity as criteria (see Table 6). They summarize by, "Values less than
1 have been obtained in areas of heavy pollution, values from 1 to 3 in areas
of moderate pollution and values exceeding 3 in clean water areas".
nn _
In a similar vein, Prophet and Edwards examined d (diversity per individual)
for macroinvertebrates in a Great Plains stream receiving feedlot runoff.
Their data are presented on Table 7. From the range of values for the 1968-
1969 period, they concluded that the system was experiencing moderate en-
vironmental stress. Cottonwood Falls and Soden's Grove were points receiving
major feedlot runoff. Statistical analysis of individual d~'s detected sig-
nificant (0.05) differences between cf's at Elmdale and West Emporia, Kansas,
(clean stations) when compared to the other three stations. Further, they
concluded that id's from the second sampling period, 1970-1971, indicated
recovery after the runoff was reduced.
21
-------
Table 6. EXAMPLES OF SPECIES DIVERSITY, d_, IN POLLUTED WATERS
(after Wilhm and Dorris, 1968)
INJ
ro
Area
Skeleton Creek
Skeleton Creek
Otter Creek
Refinery ponds
Keystone Reservoir
Alamltos Bay
Alamitos Bay
Alamitos Bay
Above
Pollutants outfall
Domestic, oil refinery *
Domestic, oil refinery 3.75
Oil field brines 3.36
Oil refinery *
Dissolved solids
Oil field brines *
Oil field brines *
Storm sewer *
d_
Near
outfall
0.84
0.94
1.58
0.98
0.55
1.49
1.44
1.45
Downstream
1.59
2.43
2.79
2.50
2.70
2.81
3.44
3.80
3.84
3.17
3.01
*
*
*
Data not available.
-------
Table 7. COMPARISON OF MEAN ANNUAL d
(DIVERSITY PER INDIVIDUAL) VALUES PER STATION
(after Prophet and Edwards, 1973)
Mean 3[
Station 1968-1969 1970-1971
Elmdale 2.86 2.73
Cottonwood Falls 2.05 3.09
West Emporia 2.70 3.18
Soden's Grove 2.02 2.57
Neosho Rapids 2.39 2.85
23
-------
Figure 1. DIT for six stations in the New River, Va. (R-rock crusher,
T-tannery, TF=textile fiber plant.) (Modified from Cairns
and Dickson, 1971.)
/6
8
L = left bank
M = middle channel
R = right bank
Ln R f LMR
I R I
LMR. 4
3 T
LHR
4
LMR. 4 LMR.
S TF 6
STATIONS
24
-------
Some of the more recent Investigators in freshwater streams have attempted
to simplify the process of obtaining diversity indices so that they may be
computed by persons without biological training. Perhaps the most elaborate
29 8 30
scheme has been devised by Cairns and others , Cairns and Dickson ' who
have devised a "sequential comparison index (S.C.I.). Data are acquired from
samples by sorting the individuals into groups with obvious differences in
appearance. Densities for the groups are obtained by counting. A sample
for a given station is randomized by gentle shaking of the collection jar
and then pouring the contents into a flat white enamel pan. Only two speci-
mens need be compared at a time. If the specimen nearest the first examined
1s similar to the first, then it is part of the same "run". If not, a new
run is begun. If fewer than 250 Individuals are in the sample, then the
Index will simply be the number of "runs" divided by the number of specimens.
If there are greater than 250 specimens, then increments of 50 specimens
are counted and an index computed as number of "runs" divided by 50. Cumu-
lative indices are calculated in increments of 50 until the plot of index
8 30
against number of specimens becomes asymptotic. Cairns and Dickson
summarize the technique in some 19 steps which should be consulted for de-
tails. They report that healthy streams with high diversity and a balanced
density seem to have DIj values above 12.0, polluted communities with skewed
population structures have given values of 8.0 or less, and that intermediate
values have been found in semipolluted situations. They present a case
history to Illustrate the usefulness of the index. Data presented on
o
Figure 1 are from Cairns and Dickson .
31
A second approach toward simplification has been taken by Egloff and Brakel .
18
They computed Patten's Index:
cT = E N./N Iog2 iyN,
where N = the total number of individuals and N. = the number of individuals
1n the 1th species. However, they substitute higher taxa for numbers of
species. Thus genera, orders, and classes are used in place of specific
identifications. Further, they compute an evenness component, from Pielou :
e = d/log2 S. Data are presented graphically on the indices computed at
the generic level for comparison of samples collected with Surber and Ekman
samplers at stations above and below a wastewater outfall. Sharp declines
25
-------
in d calculated for the total samples and for the Surber samples are evident
from a station above the outfall and one immediately below. A similar trend
is apparent for numbers of species, £, and evenness, e_. Significant (0.01)
correlations were found between d" and BOD, PO,, and ML. Significant corre-
lations to the same water quality components were detected for numbers of
species, :S.
32
Archibald compares data calculated for several common indices of species
diversity relating to South African stream diatom associations. The indices
compared are:
m 2 33
1. Simpson's index - S.I. = z IT (Duffy ),
i=l i
where IT. is the proportion in the ith species in sample.
2. Menhinick's index - S/yN~(Wilhm34),
where S is the number of species and N = total number of individuals.
3. Margalef's index - S-l/logeN (Wilhm34),
where S and N are as above.
_ m
4. Brillouin's index - H = K (log N! - z log n.!/N,
i=l n
where K , the conversion factor log-jg to logg is 3.322, N = total number of
individuals in the sample, n. = number of individuals of the ith species.
5. Patten's redundancy - R = ^-H/WW
where H = K (log N! - z log n.!), and H. = K {log N! - [log N-(S-l)]!},
• n i min
where S = number of species in sample; N and n. are as above.
6. Cairn's sequential comparison index from:
"runs"/200 where 200 specimens were used in each case.
Resulting data are presented on Table 8. Table 9 is a listing of the corre-
lation coefficients.
32
From the correlation data, Archibald concludes that all pairs of indices
have significant correlation (P = 99.9 percent), and thus selection of an
index for use should be based on three considerations, i.e., (1) time re-
quired to sort the-sample; (2) ease of index calculation; and (3) the
characteristic of fixed limits. Archibald concluded the S.C.I, best met
each of the criteria.
26
-------
Table 8. SPECIES DIVERSITY VALUES FOR THE SAMPLES UNDER OBSERVATION
AS OBTAINED FROM DIFFERENT DIVERSITY INDICES
(after Archibald, 1972)
Diversity Indices
Samp] e
number
306
315
324
338
339
340
342
344
350
353
375
377
406
464
493
495
497
498
K22
s
11
38
41
20
41
30
39
43
23
32
14
12
28
16
22
27
14
17
15
N
441
399
394
382
410
372
366
386
462
399
375
364
387
388
392
423
450
396
442
S.I.
0.82
0.08
0.10
0.16
0.07
0.12
0.09
0.10
0.54
0.27
0.33
0.79
0.28
0.58
0.41
0.23
0.70
0.20
0.74
s/ N
0.524
1.903
2.065
1.024
2.025
1.555
2.039
2.188
1.070
1.602
0.728
0.629
1.423
0.869
1.111
1.313
0.660
0.854
0.715
s - 1/lnN
1.64
6.18
6.69
3.20
6.65
4.90
6.44
7.05
3.59
5.18
2.19
1.87
4.53
2.52
3.52
4.30
2.13
2.67
2.30
H
0.638
4.023
3.777
3.026
4.135
3.526
3.872
3.822
1.715
2.928
1.950
0.749
2.348
1.457
2.106
2.688
1.087
2.717
0.959
R
0.862
0.196
0.283
0.307
0.221
0.305
0.294
0.307
0.673
0.466
0.506
0.849
0.557
0.690
0.565
0.447
0.760
0.352
0.810
S.C.I,
0.320
0.840
0.885
0.780
0.885
0.845
0.910
0.905
0.615
0.855
0.700
0.435
0.665
0.550
0.800
0.725
0.235
0.690
0.165
-------
Table 9. CORRELATION COEFFICIENTS OF DIFFERENT PAIRS OF DIVERSITY INDICES3
(after Archibald, 1972)
S.C.I.
S.I.
S/v^N
S - l/logeN
H
R
-0.89
+0.98
-0.81
-0.76
-0.98
H
+0.90
-0.97
+0.91
-0.81
S - I/log N S/v'N
e
+0.80 +0.81
-0.82 -0.89
+0.99
S.I.
-0.92
aWith 17 degrees of freedom p > 99.9 per cent (sic) when r_ > 0.6932 (Fisher
and Yates (1948)-Statistical Tables for Biological and Medical Research.
Oliver & Boyd, Edinburgh, Table VI).
28
-------
Archibald applied S.C.I, to diatom associations with an Interesting and In-
formative result. The data, presented in graphical form, Figure 2 (A through
D) show differing values for the index but not in relation to the "polluted"
nature of the areas from which the collections were made. Thus, Figure 2A
represents a "clean" water association with a S.C.I, value of 0.725; Figure
2C is from a moderately stressed region having a slightly higher S.C.I, value,
0.885. Figure 2C Is an association from "clean" water which is dominated by
a single species with an S.C.I, value of 0.550 while Figure 2D is an associ-
ation from a heavily polluted region having an S.C.I, value of 0.165.
Archibald recognizes that the diatom associations may be quite different from
the macroinvertebrate "communities" for which good correlations between in-
dex values and pollution have been found. He does, however, point out that
the results of his study serve to illustrate the need for caution in inter-
pretation of raw index values in pollution studies.
Further support for the use of such caution in the use of the information
index alone comes from Lotrich who used H to detect the effects of strip
mine wastes on stream fishes. He determined that abundance of the fishes
was reduced by about one-half but that since the reduction was somewhat
equally distributed among the species an effect ws not elucidated by the
index. Dickman in working with algae, found that an index based on pro-
ductivity was more useful than the Shannon information index. Whiteside and
McNatt obtained a positive correlation of the information index for stream
fishes to stream order. Inexplicably, however, the relationship did not
hold for the highest stream order. The authors attributed the exception to
sampling inefficiency.
38
McKay and Kalft used a species richness type index, d = S-l/log N, to
evaluate small stream benthos. They determined statistically significant
differences in seasonal values with higher values for the index in winter
and in summer than for fall and spring.
Several authors have conducted experimental studies in laboratory streams
39
using the Shannon information index. Mitchell used freshwater algae in
an experimental situation. He relied on the index to determine the effects
of the addition of aliquots of wastewater and detergents and concluded from
index values that the contaminants produced no change. In experimental
29
-------
Figure 2. Structure of diatom communities (modified from Archibald,*1972)
JO-
£0-
» to -
•k
population
1 Fig. 2A
L 495
S.C.I.=0.725
v
l^
5 to tS
Fig. 2C
339
5.C.I.=0.885
I I I I M I M I M I
to
o>
(U
•r«
U
V
CL
o
70 -
60-
40*
40-
/O-
Fig. 2B
464
S.C.I. =0.550
I ) t I T 1 I I I III I I I
to ts
Fig. 20
BK 22
S.C.I.=0.165
I I II 'l"l"l I M I I 1 1 I
5 ttt tS
Species number
30
-------
40
streams, Ewing and Dorris found an increasing value for the information
index, F through time, but found no significant correlation to the P/R ratio.
41
Likewise Kehde determined increasing diversity values through time in
laboratory streams but did not detect a difference in diversity for grazed
42
versus ungrazed periphyton. Patrick obtained similar values (about 3.2)
for the Shannon index in freshwater stream replicates of similar character.
TERRESTRIAL
Unlike the situation in freshwater, investigators in the terrestrial habitat
have emphasized the use of diversity indices to describe the effects of
natural variables, i.e., food density, precipitation, cover, predation.
Further, in the terrestrial habitat there has been a much greater distinction
between the components of diversity than was the case for many freshwater
studies. There has been much less tendency to ascribe specific levels for
a particular index in the terrestrial habitat. There is an apparent increase
in diversity with community succession.
Information theory type indices have been used most frequently. The dis-
cussion of bird species diversity in terms of foliage has received attention.
MacArthur and MacArthur proposed niche description for birds in terms of
AA
the diversity index, BSD = -zpi log P... MacArthur found that BSD relatad
positively to foliage height diversity within habitats but not across habi-
tats. Further, he pointed out that 20 to 25 pairs of birds were needed for
comparison and that sampling schemes based on a specific area might not pro-
45
vide an adequate sample. Karr examined the distribution of Shannon's H
with respect to birds in lowland tropical grass, shrub, and forest habitats.
Diversity, as indicated by the index, was higher for shrubs than grass or
forest and was relatively stable seasonally. Kricher used an information
H' and the J1 component of Pielou for evenness of distribution. He deter-
mined a low and variable value to be characteristic of early successional
stages or ecosystems characterized by opportunistic species. The informa-
tion index, H1, more nearly correlated to number of species than did the
evenness component. J1 distinguished nesting and territorality since it was
47 '
stable from census to census. In a later paper Kricher gave some time
values for development of diversity. Thus, he states that bird species
diversity on a developing sere increases with time to about 150 years.
31
-------
48
The BSD-time curve flattens somewhat after about 30 years. Wiens used
49
the information Index and the equltabillty component of Lloyd and Ghelardi
for birds in forests. He found that the values were uniformly low and that
neither index gave a consistent pattern with habitat heterogeneity.
50
Turning to other species in the terrestrial habitat, Luff found the
49
equitability index of Lloyd and Ghelardi to be useful in describing the
distribution of the beetle fauna of grass tussocks. He limited interpretive
51
remarks to a discussion of his own data. Monk and others observed a pos-
itive relationship between MacArthur's index, -z p. log P., and sample size
52
in an oak-hickory forest. Fleming used the information index, H1, to
describe the world distribution of terrestrial mammals. He determined a
53
southward increase in numbers of species. Heyer and Berren described the
distribution of frogs, lizards, and snails of tree buttresses in Thailand
and Equador. A higher diversity was found in Equador. To illustrate the
54
versatility of the information index, Hurtubia compared the food of lizards
55
to lizard diversity. Brown found the information index to be well corre-
lated to predictable annual rainfall when applied to sand dune rodents.
Coulson compared beetle diversity in monocultures of white pine to that
in mixed coppice stands and determined higher diversity in the mixed stands.
Murdock and others used the Brillouin information index and obtained a
good correlation between insect diversity and plant diversity. They ob-
Sfi
served a consistent midsummer dip in index values. Randolph obtained
values of 0.76 in Johnson grass to 1.57 in woods for the Shannon index. He
concluded that the relative abundance component was more influential than
5
numbers of species. Pianka used the information index, H, in comparing
lizards of North American deserts. He concluded that ecological time would
be Important in determining diversity only when there were pronounced
barriers to dispersal. He further concluded that numbers of species, as an
indicator, would only be useful as a long-term parameter, i.e., greater than
59
five years. Shafi and Yarranton studied the effects of fire on the suc-
cession of a forest. In concluding remarks they state, "The weakness of
diversity as an ecological tool lies in its ambiguity". They further point
out that one must always consider at least two components. The long-term
successional trend was 1n the direction of declining diversity values. They
32
-------
attributed high values and fluctuations during four to eleven years after
burnings to the effects of species richness rather than evenness.
In an experimental study of terrestrial grasslands, Mai one used the species
richness formula, d = S-1/log.N, to detect the effects of an arsenical herbi-
—• g
cide. He found that d_ declined in proportion to the amount of chemical added.
MARINE
In the marine environment, diversity indices used have mostly been those
based on information theory although there has been a considerable impact
19
in the literature based on Sanders rarefaction techniques for determining
the effects of differing sample sizes. Information type indices were used
by Jackson , Abele , Boesch , Cameron , Cooper and Copeland , Coull ,
Dahlberg and Odum , Johnson and Brinkhurst , Johnson ', Kohn , Lie
and Kisker72, Lie and Evans73, Patten74, and Porter75'76'
Investigators of the marine environment who have used the information index
62
have been predominantly concerned with the marine benthos. Abele computed
the information index, H1, and the evenness component, J1, (Lloyd and
49
Ghelardi ) for marine decapods. He found good correlation of H1 and sedi-
ment type but did not find a significant correlation of the index to temper-
ature or tidal exposure. Jackson compared animal diversity on Thallassia
beds in the intertidal zone of a bay and exposed coast. He found a greater
diversity, as revealed by Brillouin's index in the bay habitat. Johnson
determined an increasing value of H (Brillouin) from high intertidal to sub-
tidal. He introduced a ranking system based on the order of appearance in
succession to a stress gradient. In a later paper Johnson states that
polychaete and mollusk communities are sensitive to environmental change.
72
Lie and Kisker determined increasing diversity with an information index
from shallower to deeper subtidal waters applied to the benthos. They
postulated that the trend was related to greater environmental stability.
Lie and Evans discussed the variability in the information index under
natural or baseline conditions. Coull used several indices to compare
marine microbenthos and concluded that the Shannon information index 'agreed
19
well with Sanders rarefaction techniques and that the indices indicated
that diversity increased with depth and environmental stability. Boesch
compared the Shannon information index and others for analysis of estuarine
33
-------
benthos and concluded that only If several Indices were used concurrently,
useful information could be obtained.
Information theory indices have also been applied to estuarine fish, plankton,
marine corals, and salt marshes. Dahlberg and Odum applied several indices
to estuarine fish and concluded, "In terms of practical application of di-
versity indices to detection and evaluation o? pollution, it would appear
that indices H" and D are seasonally stable and, therefore, suitable general
indices that could be applied to any season". Further on they conclude,
"It is evident that a combination of indices which reflect the different
components of diversity should be selected and that these should be based
on the seasonal and sample characteristics of populations to be monitored".
74 m
Patten developed a diversity index, D = E X. log,, (X./X), to describe
i=l i * i
diversity of marine plankton. He found that diversity dropped off abruptly
at compensation depth. Maximum diversity occurred at 60 cm depth and Patten
related his index to system energy. Cooper and Copeland constructed a
model of Galveston Bay to which they applied an information theory index
for zooplankton. They determined a positive correlation between index value
and system development time and a reduced index value in response to 15
percent industrial effluent.
Cameron studied the relationship of the Brillouin information index to
physical and chemical variables in a salt marsh. He states, "Physical micro-
environmental factors, especially temperature and vapor pressure deficit,
seemed to be important in cuing larval development, but did not exert a
dramatic effect on adult diversity trends". His study included multivariate
analysis and although significant microenvironmental effects on the index
were determined, the analysis did not detect spraying with insecticide as a
significant effect on insect diversity.
Porter * studied marine corals in terms of an information theory index.
He concluded that a high concentration of predators produced a higher di-
versity in coral and attributed the higher diversity to the evenness com-
ponent. He states, "...best single indicator of species diversity is ^".
(Number of species.)
At least two authors have used species richness type indicators for marine
plankton analysis. Ignatiades used the formula, D = 1/N log, N!/N ,, for
& as
34
-------
pelagic plankton and concluded, "Diversity .can never be computed on a total
community". He further Indicates, d declines with "blooms" and increases
75"
with a decline in "blooms". Hardy used a species richness index and de-
termined that the index fluctuated with season giving high values in summer
and low values in winter. Further, Hardy reports that high productivity
gave low diversity.
19
In an effort to compare effects of sample size, Sanders developed a
technique for estimating diversity from different sizes of samples and com-
paring the estimates. He discusses the necessity of comparing only for com-
parable habitats, In his case soft mud substrates from different parts of
the world. He concluded that the Shannon-Wiener function (information index)
was relatively free from sample size problems. Sanders reported an increase
79
in diversity with increasing depth, as determined by rarefaction. Gage ,
80 81
Dexter , and Day have used the rarefaction methodology (graphical) in
similar marine benthic situations and have obtained predictable results.
82
Simberloft pointed out that the rarefaction technique overestimates the
number of species.
COMPARATIVE INDEX STUDIES
Several authors have conducted studies particularly for the purpose of com-
paring various species diversity indices. As pointed out in the section on
32
freshwater, Archibald compared a number of common indices, determined a
high correlation between indices, and concluded an index from those tested
should be chosen based on ease of application. He suggested Cairns and
29 8
others sequential comparison index. Cairns and Dickson suggested the
use of the sequential comparison index for untrained persons but recommended
22
that work continue on the information theory derived indices. Eberhardt
considered Indices and concluded that any one of several mathematical dis-
tributions were adequate but that sampling data had not been emphasized.
Further, he concluded that the various indices could serve as a convenient
83
method of summarization but not as predictive tools. Loya compared
species counts, Simpson's Index, and several information theory-derived in-
dices and pointed out that while the diversity of hematypic corals increased
with depth as measured by the several indices, there was a great need for
the use of multiple indices. Mills stated that we "...cannot define
community rigorously...," it is "...an ecological unit of any degree".
35
-------
He concluded that it was much too early to use diversity indices as a
85
quantitative tool. Turner and Broadhead used Fisher's index, Mclntosh's
index, and Brillouin's index for microepiphytes of the bark surface and Ob-
oe
tained similar results with each of them. DeBenedictis in a study con-
sidering species richness indices and information theory derived indices
with evenness components, concluded that the correlations obtained were
correlations of mathematics only. He stated that a relationship between
any of the indices and a biological phenomenon was lacking. Hill presents
a unified notation which claims to interrelate several of the more common
13
indices of species diversity as a continuum. Hurlbert says that species
diversity is a non-concept in ecology.
36
-------
SECTION VI
REFERENCES
1. Woodwell, G. M. and H. H. Smith (eds). Diversity and Stability in
Ecological Systems. Brookhaven National Laboratory Publication
No. 22, Upton, N. Y. 1969. 264 p.
2. Connell, J. H. and E. Orias. The Ecological Regulation of Species
Diversity. Am Nat. 98:399-414, 1964.
3. Whittaker, R. H. Dominance and Diversity in Land Plant Communities.
Science . J47 (3655):250-260, 1965.
4. MacArthur, R. H. Patterns of Species Diversity. Biol. Rev. 4():
510-533, 1965.
5. Pianka, E. R. Latitudinal Gradients in Species Diversity: A Review
of Concepts. Am Nat. 100 (910): 33-46, 1966.
6. Mclntosh, R. P. An Index of Diversity and the Relation of Certain
Concepts of Diversity. Ecology. 48: 392-404, 1967.
7. Wilhm, J. L. and T. C. Dorris. Biological Parameters for Water Quality
Criteria. BioScience. J18 (6): 477-481, 1968.
8. Cairns, J. and K. L. Dickson. A Simple Method for the Biological
Assessment of the Effects of Waste Discharges on Aquatic Bottom-
Dwelling Organisms. J Wat Poll Cont Fed. 43 (5): 755-772, 1971.
9. Warren, C. E. Biology and Water Pollution Control (in collaboration
with P. Doudoroff). Philadelphia, W. B. Saunders Co., 1971. 434 p.
10. Hill, M. 0. Diversity and Evenness: A Unifying Notation and Its Con-
sequences. Ecology. 54: 427-433, 1973.
11. Pielou, E. C. The Use of Information Theory in the Study of Biological
Populations. In: Proc 5th Berkeley Symposium on Mathematical
Statistics and Probability, Berkeley, Univ of Calif Press, 4;
163-177, 1967.
12. Wilhm,, J. L. Range of Diversity Index in Benthic Macroinvertebrate
Populations. J Wat Poll Cont Fed. 42 (5): Part 2, R221-R224, 1970.
13. Hurlbert, S. H. The Nonconcept of Species Diversity: A Critique and
Alternative Parameters. Ecology. 52 (4): 577-586, 1971.
37
-------
14. Environmental Protection Agency. Biological Field and Laboratory
Methods for Measuring the Quality of Surface Waters and Effluents.
Cincinnati, Ohio. Publication Number EPA-670/4-73-001. July 1973.
15. Fisher, R. A., A. S. Corbet, and C. B. Williams. The Relation between
the Number of Species and the Number of Individuals in a Random
Sample of an Animal Population. J Anim Ecol. ]2: 42-58, 1943.
16. MacArthur, R. H. and E. 0. Wilson. The Theory of Island Biogeography.
Princeton, Princeton Univ Press, 1967. 203 p.
17. Margalef, R. Information Theory in Ecology. General Systems. 3; 36-
71, 1958.
18. Patten, B. C. Species Diversity in Net Plankton of Raritan Bay. J
Mar Res. 20: 57-75, 1962.
19. Sanders,. H. L. Marine Benthic Diversity: A Comparative Study. Am Nat.
]Q2 (925): 243-282, 1968.
20. Shannon, C. E. and W. Weaver. The Mathematical Theory of Communication.
Urbana, Univ of Illinois Press, 1968. 117 p.
21. Simpson, E. H. Measurement of Diversity. Nature. 163: 688. 1949.
22. Eberhardt, L. L. Some Aspects of Species Diversity Models. Ecology.
50 (3): 503-505, 1969.
23. Gilbert, E. N. Information Theory after 18 Years. Science. 152; 320-
326, 1966.
24. Margalef, R. Diversidad de Species en las Comunidades Naturales.
Publicaciones Institute de Biologia Aplicada. 9; 5-28, 1951.
25. Informacion y Diversidad Especifica en las Comunidades de
Organisms. Invest Pesq. 3; 99-106, 1956.
26 Brillouin, L. Science and Information Theory. New York, Academic Press,
1960.
27. Wilhm, J. L. and T. C. Dorris. Species Diversity of Benthic Macro-
invertebrates in a Stream Receiving Domestic and Oil Refinery
Effluents. Am Mid Nat. T£ (2): 427-449, 1966.
38
-------
28. Prophet, C. W. and N. L. Edwards. Benthic Macroinvertebrate Community
Structure in a Great Plains Stream Receiving Feedlot Runoff.
Water Res Bull Amer Water Res Ass. i (3): 583-589, 1973.
29. Cairns, J., D. W. Albaugh, F. Busey, and M. D. Chanay. The Sequential
Comparison Index: A Simplified Method for Non-Biologists to Estimate
Relative Differences in Biological Diversity in Stream Pollution
Studies. J Water Poll Cont Fed. 40 (9): 1607-1613, 1968.
30. Cairns, J. and K. L. Dickson, eds. Biological Methods for the Assess-
ment of Water Quality. Philadelphia, ASTM Special Publication
528, 1973. 256 p.
31. Egloff, D. A. and W. H. Brakel. Stream Pollution and a Simplified
Diversity Index. J Wat Poll Cont Fed 45 (11): 2269-2275, 1973.
32. Archibald, R. E. M. Diversity in Some South African Diatom Associations
and its Relation to Water Quality. Water Res. B (10): 1229-1238,
1972.
33. Duffy, E. An Ecological Analysis of the Spider Fauna of Sand Dunes.
J Anim Ecol. 37: 641-674, 1968.
34. Wilhm, J. L. Comparison of Some Diversity Indices Applied to Popula-
tions of Benthic Macroinvertebrates in a Stream Receiving Organic
Wastes. J Wat Poll Cont Fed. 39: 1673-1683, 1967.
35. Lotrich, V. A. Growth, Predation, and Community Composition of Fishes
Inhabiting a First, Second, and Third Order Stream of Eastern
Kentucky. Ecol Monogr. 43 (3): 377-397, 1973.
36. Dickman, M. Some Indices of Diversity. Ecology. 49_ (6): 1191-1193,
1968.
37. Whiteside, B. G. and R. M. McNatt. Fish Species Diversity in Relation
to Stream Order and Physicochemical Conditions in the Plum Creek
Drainage Basin. Am Mid Nat. 88: 90-101, 1972.
38. McKay, R. J. and J. Kalft. Seasonal Variation in Standing Crop and
Species Diversity of Insect Communities in a Small Quebec Stream.
Ecology. 50 (1): 101-109, 1969.
39
-------
39. Mitchell, D. Eutrophication of Lake Water Microcosms: Phosphate versus
Nonphosphate Detergents. Science. 174 (4011): 827-829, 1971.
40. Ewing, M. S. and T. C. Dorris. Algal Community Structure in Artificial
Ponds Subjected to Continuous Enrichment. Am Mid Nat. 83_ (2):
565-582, 1970.
41. Kehde, P. M. and J. L. Wilhm. The Effects of Grazing by Snails on
Community Structure of Periphyton in Laboratory Streams. Am Mid
Nat. 87 (1): 8-24, 1972.
42. Patrick, R. The Structure of Diatom Communities in Similar Ecological
Conditions. Am Nat. J02 (924): 173-183, 1968.
43. MacArthur, R. H. and J. W. MacArthur. On Bird Species Diversity.
Ecology. 42 (3): 594-598, 1961.
44. MacArthur, R. H. Environmental Factors Affecting Bird Species Diversity.
Am Nat. 98 (903): 387-397, 1964.
45. Karr, J. R. Structure of Avian Communities in Selected Panama and
»
Illinois Habitats. Ecol Monogr. 41_ (3): 207-233, 1971.
46. Kricher, J. R. Summer Bird Species Diversity in Relation to Secondary
Succession on the New Jersey Piedmont. Am Mid Nat.89_ (1): 121-
137, 1973.
47. Bird Species Diversity: The Effect of Species Richness
and Equitability on the Diversity Index. Ecology. 53_ (2): 278-
282, 1972.
48. Weins, J. A. Habitat Heterogeneity and Avian Community Structure.
Am Mid Nat. 91_ (1): 195-213, 1974.
49. Lloyd, M. and R. J. Ghelardi. A Table for Calculating the Equitability
Component of Species Diversity. J Anim Ecol. 33_: 421-425, 1964.
50. Luff, M. L. The Abundance and Diversity of the Beetle Fauna of Grass
Tussocks. J Anim Ecol. 35 (1): 189-208, 1966.
51. Monk, C. D., C. I. Child, and S. A. Nicholson. Species Diversity of a
Stratified Oak-Hickory Community. Ecology. J50 (3): 468-470, 1969.
52. Fleming, T. H. Numbers of Mammal Species in North and Central American
Forest Communities. Ecology. 54 (3): 555-563, 1973.
40
-------
53. Heyer, W. R. and K. A. Berren. Species Diversities of Harpetofaunal
Samples from Similar Microhabitats at Two Tropical Sites. Ecology.
54 (3): 642-645, 1973.
54. Hurtubia, J. Trophic Diversity Measurements in Sympatric Predatory
Species. Ecology. 54 (4): 885-890, 1973.
55. Brown, J. H. Species Diversity of Seed-Eating Desert Rodents on Sand
Dune Habitats. Ecology. 54 (4): 775-787, 1973.
56. Coulson, R. N., D. A. Crussley, Jr., and C. S. Gist. Patterns of
Coleoptera Species Diversity in Contrasting White Pine and Coppice
Canopy Communities. Am Mid Nat. 86 (1): 145, 1971.
57. Murdock, W. W., F. C. Evans, and C. H. Peterson. Diversity and Pattern
in Plants and Insects. Ecology. 53: 819-829, 1972.
58. Randolph, P. A. Influence of Environmental Variability on Land Snail
Population Properties. Ecology. 54 (4): 933-955, 1973.
59. Shafi, M. I. and G. A. Yarranton. Diversity, Floristic Richness, and
Species Evenness Ouring a Secondary (post-fire) Succession. Ecology.
54 (4): 897-902, 1973.
60. Malone, C. R. Effects of a Non-selective Arsenical Herbidice on Plant
Biomass and Community Structure in a Fescue Meadow. Ecology. 53_:
507-512, 1972.
61. Jackson, J. B. C. The Ecology of the Molluscs of Thallassia Communities,
Jamaica, West Indies. II. MolTuscan Population Variability along
an Environmental Stress Gradient. Mar Biol. ^4 (4): 304-337, 1968.
62. Abele, L. G. Species Diversity .of Decapod Crustaceans in Marine
Habitats. Ecology. 55 (1): 156-161, 1974.
63. Boesch, D. F. Classification and Community Structure of Macrobenthos
in Hampton Roads Area, Virginia. Mar Biol. 21 (3): 226-244, 1973.
64. Cameron, G. N. Analysis of Insect Trophic Diversity in Two Salt Marsh
Communities. Ecology 53 (1): 58-73, 1972.
65. Cooper, D. C. and B. J. Copeland. Responses of Continuous - Series
Estuarine Microorganisms to Point - Source Input Variations.
Ecol Monogr. 43: 213 -236, 1973.
41
-------
66. Coull, B. C. Species Diversity and Fauna! Affinities of Meiobenthic
Copepoda in the Deep Sea1. Mar Biol. 14. (1): 48-51, 1972.
67. Dahlberg, M. D. and E. P. Odum. Annual Cycles of Species Diversity
in Georgia Estuarine Fish Populations. Am Mid Nat. 86_ (1): 145,
1971.
68. Johnson, M. G. and R. 0. Brinkhurst. Associations and Species Diversity
1n Benthic Macroinvertebrates of Bay of Quinte and Lake Ontario.
J Fish Res Board Can. 28 (11): 1683-1697, 1971.
69. Johnson, R. G. Variations in Diversity within Benthic Marine Communities.
Am Nat. ]04 (937): 285-300, 1970.
70. Animal-Sediment Relations in Shallow Water Benthic
Communities. Mar Geol. IT. (2): 93-104, 1971.
71. Kohn, A. J. Environmental Complexity and Species Diversity in the
Gastropod Genus Conus on Indo-West Pacific Reef Platforms. Am Nat.
101 (919): 251-259, 1967.
72. Lie, U. and D. S. Kisker. Species Composition and Structure of Benthic
Infauna Commun+ties off the Coast of Washington. J. Fish Res.Board
Can. 27 (12): 2273-2285, 1970.
73. and R. A. Evans. Long-term Variability in the Structure of
Subtidal Benthic Communities in Puget Sound, Washington, U.S.A.
Mar Biol. 21. (2): 122-126, 1973.
74. Patten, B. C. Plankton: Optimum Diversity Structure of a Summer
Community. Science J40 (3569): 894-898, 1963.
75. Porter, J. W. Predation by Acanthaster and its Effect on'Coral Species
Diversity. Am Nat. J06 (950): 487-492, 1972a..
76. Patterns of Species Diversity in Caribbean Reef Corals.
Ecology 53 (4): 745-748, 1972b..
77. Ignatiades, L. Annual Cycle, Species Diversity and Succession of
Phytoplankton in Lower Saronicos Bay, Aegean Sea. Mar Biol. 3_ (3):
196-200, 1969.
42
-------
78. Hardy, J. T. Phytoneuston Ecology of a Temperature Marine Lagoon.
Umnol Oceanogr. 1JJ (4): 525-533, 1973.
79. Gage, J. Community Structure of the Benthos in Scottish Sea-Lochs.
I. Introduction and Species Diversity. Mar Biol. 20 (2): 89-100,
1973.
80. Dexter, D. M. Structure of an Intertidal Sandy-Beach Community in
North Carolina. Chesapeake Sci. 1_0 (2): 93-98, 1969.
81. Day, J. H., J. G. Field, and M. P. Montgomery. The Use of Numerical
Methods to Determine the Distribution of Benthic Fauna across
the Continental Shelf of North Carolina. J Anim Ecol. 40 (1):
93-125, 1971.
82. Simberloff, D. Properties of the Rarefaction Diversity Measurements.
Am Nat. J06 (949): 414-418, 1972.
83. Loya Y. Community Structure and Species Diversity of Hermatypic Corals
at Eilat, Red Sea. Mar Biol. 1_3 (2): 100-123, 1972.
84. Mills, E. L. The Community Concept in Marine Zoology, with Comments on
Continua and Instability in Some Marine Communities: A Review.
J Fish Res Board Can. 26 (6): 1415-1428, 1969.
85. Turner, B. 0. and E. Broadhead. The Diversity and Distribution of
Psocid Populations on Mangifera idica L. in Jamaica and Their Re-
lationship to Altitude and Macroepiphyte Diversity. J Ecol. 62
, 1974.
86. DeBenedictis, P. A. On the Correlations between Certain Diversity
Indices. Am Nat. 107 (954): 295-302, 1973.
43
-------
APPENDIX A
LETTER OF INQUIRY USED TO DETERMINE UNDERSTANDING OF COMMUNITIES
45, 46, 47
-------
COMMUNITY PARAMETERS FOR ENVIRONMENTAL ASSESSMENT
Battelle-Northwest, in conducting a study for the U.S. Environmental Protection
Agency, is attempting to ascertain if biological community dynamics common to the marine,
freshwater, and terrestrial habitats are useful in the assessment of environmental change.
As a part of these studies, we are seeking an opinion from the industrial sector, from regula-
tory agencies, and from researchers representing the relevant technical expertise. The
following questions are designed to develop information regarding the understanding and
acceptance of community response as a measure of environmental change. Please return
completed inquiries in the postage paid envelopes provided. Replies should be anonymous.
SECTION I — IDENTIFICATION OF RESPONDENT
1. Of the three categories mentioned above, I am most nearly associated with:
D Industry D Regulatory agency a Research
2. With respect to interest in environmental change, my main involvement is with the (one or more):
D Terrestrial n Freshwater D Marine
3. My participation in environmental studies has been:
a None a Occasional o Frequent
4. My understanding of biological communities is:
a Limited n Moderate Q Comprehensive
5. Position
Years of experience in environmental assessment
D Fulltime
SECTION II — TYPE OF INVOLVEMENT
1. My participation in environmental studies has principally been with:
D Chemical criteria n Laboratory toxicity studies
D Other (specify)
n Field surveys
2. I have had occasion to deal with community parameters or indices of species diversity for
environmental assessment:
a Never D Occasionally Q Frequently
3. Please indicate the habitat for which you have applied the following indices:
Terrestrial
Margalef d~= (S-1)/logeN
Marine
D
Brillouin
(logN! -ElogNjI)
Wilhm and Dorris modification of Brillouin
d = E(n,/n) logj(nj/n)
Patrick Aquatic community indices of pollution
Wurtz Modification of Patrick method (tolerance)
D
a
D
D
a
a
-------
SECTION II — TYPE OF INVOLVEMENT (CONTINUED)
Fisher
Simpson
ax+ox 2/2 = ox 3/3+.. . +ax h/h
NVtn(n-l)
2j
I = 2ab~(a+b)
Mountford
Lloyd, Zar, and Kar d = C/N (N logi0N - ni Iog10ni)
Beck Biotic index = 2(n Class I) + (Class II)
Aerial Surveys
Experimental plots (quadrats)
Other (please specify)
Terrestrial
D
D
a
D
D
D
D
O
Freshwater
Q
O
D
0
a
o
a
Q
Marine
D
D
D
D
a
a
D
4. Comments:
SECTION III — OPINION
1. In your opinion, an indicator of environmental change should be mainly:
a Chemical analyses D The organism a The population
o The community D Combination of above (please specify)
Q No opinion
o Other (please specify)
2. Is there a parameter of community response (index) indicative of environmental change?
D None o One D More than one o No opinion
3. Would you recommend any of the indices listed in Section II for general use by personnel with
limited training?
o No
o Yes (which one(s))
4. Are there other indices or methods you have used to measure environmental change?
D No D Yes (please specify)
Comments
5. In general, how do you rate the usefulness of community analysis?
D Poor o Fair D Good a Excellent
Would you participate in a personal interview with one of our investigators?
o Yes (please send telephone number or call collect, 206/683-4151)
O No
Peter Wilkinson
Battelle-Northwest Marine
Research Laboratories
Rt. 2, P.O. Box 1421
Sequim, WA 98382
206/683-4151
-------
APPENDIX B
DATA SUMMARY OF WRITTEN INQUIRIES BY ASSOCIATIVE GROUP
Section I. Identification of respondent
Industry Regulatory Agency
Question Number
2. HABITAT
Terrestrial
Freshwater
Marine
Unidentified
Total
3. PARTICIPATION
None
Occasionally
Frequently
Full time
Inconclusive
Total
4. UNDERSTANDING
Limited
Moderate
Comprehensive
Inconclusive
Total
5. a. POSITION
Biologist-ecologist
Chemist
Engineer
Enforcement officer
Other
Unidentified
Total
7
10
5
7
29
1
7
9
12
0
29
8
13
7
1
29
9
1
1
0
17
1
29
Percent
24
34
17
25
100
3
24
31
42
0
100
28
45
24
3
100
31
3
3
0
59
4
100
Number
7
56
7
13
83
4
15
31
32
1
83
4
31
44
4
83
35
2
14
4
27
1
83
Percent
8
67
8
17
100
5
18
37
39
1
100
5
37
53
5
100
42
2
17
5
33
1
100
Research
Number
114
67
18
23
222
3
55
95
63
6
222
10
80
132
0
222
151
1
1
0
10
59
222
Percent
51
30
8
11
100
1
25
43
28
3
100
5
36
59
0
100
68
0
0
0
5
27
100
48
-------
APPENDIX B (continued)
Industry
Question Number Percent
5. b. EXPERIENCE (years)1
0 - 2
2 - 5
5 - 10
10+
Inconclusive
Total
2
6
8
13
0
29
7
21
27
45
0
100
Section II.
1. CRITERIA USED
Chemical
Laboratory toxicity
Field survey
Other
Inconclusive
Total
2. PARTICIPATION
Never
Occasionally
Frequently
Inconclusive
Total
3. INDICES USED
Terrestrial
Freshwater
Marine
Total
5
1
9
8
6
29
10
13
5
1
29
13
11
6
19
Percent respondents
using index on one
or more habitats
17
3
31
28
21
100
34
46
17
3
100
66
Regulatory agency
Number Percent
8
23
23
29
0
83
10
28
28
34
0
100
Research
Number
10
42
47
116
7
222
Percent
5
19
21
52
3
100
Type of involvement
10
2
43
2
26
83
10
45
23
5
83
17
59
45
70
12
2
52
2
32
100
12
54
28
6
100
84
3
6
130
16
67
222
26
113
79
4
222
133
96
34
199
1
3
59
7
30
100
12
51
35
2
100
90
1
Responses falling on a division were elevated to the next higher rank.
49
-------
APPENDIX B (continued)
Section
Industry
Question Number Percent
1 . INDICATOR
Chemical
Organism
Population
Community
Combination
No opinion
Other
Inconclusive
Total
Mean number indi-
cators/respondent
12
9
12
16
13
2
4
0
68
41
31
41
55
45
7
14
0
2.3
III. Opinion
Regulatory agency
Number Percent
28
28
26
56
51
3
1
1
194
34
34
31
67
61
4
1
1
2.3
Research
Number
62
75
116
156
134
3
13
4
563
Percent
28
34
52
71
61
1
6
2
2.5
2. NUMBER OF PARAMETERS
None
One
More than one
No opinion
Inconclusive
Total
3. LIMITED TRAINING
No
Yes
Inconclusive
Total
4. OTHER INDICES
No
Yes
Inconclusive
Total
3
1
13
11
1
29
17
2
10
29
9
11
9
29
10
4
45
38
3
100
59
7
34
100
31
38
31
100
2
0
52
15
14
83
51
18
14
83
31
35
17
83
2
0
63
18
17
100
61
22
17
100
37
42
21
100
10
2
171
17
21
22
120
63
38
221
75
103
43
221
5
1
77
8
9
100
55
28
17
100
34
47
19
100
50
-------
APPENDIX B (continued)
Industry Regulatory agency Research
Question Number Percent Number Percent Number Percent
5. RATING
Poor 5 17 45 31
Fair 4 14 19 23 42 19
Good 10 34 30 36 83 38
Excellent 3 10 15 18 57 26
Inconclusive 7 25 15 18 36 16
Total 29 100 83 100 221 100
51
-------
APPENDIX C
DATA SUMMARY OF WRITTEN INQUIRIRES BY PRINCIPAL HABITAT
Section II. Type of involvement
Terrestrial
1.
2.
3.
Question
CRITERIA USED
Chemical
Laboratory toxicity
Field survey
Other
Inconclusive
Total
PARTICIPATION
Never
Occasionally
Frequently
Inconclusive
Total
INDICES USED
Terrestrial
Freshwater
Marine
Total
Number
2
2
94
11
19
128
13
75
36
4
128
120
19
0
120
Percent
2
2
73
8
15
100
10
59
28
3
100
Freshwater
Number
11
3
55
4
58
131
13
68
45
5
131
18
114
21
117
Percent
8
2
42
3
45
100
10
52
34
4
100
Marine
Number
0
2
18
2
7
29
2
13
14
0
29
3
7
27
27
Percent
0
7
62
7
24
100
7
45
48
0
100
Undecided
Number
3
2
19
6
15
45
16
14
12
3
45
23
25
11
26
Percent
7
5
42
13
33
100
35
31
27
7
100
Percent respondents
using index on one
or more habitats
94
89
93
58
-------
APPENDIX C (continued)
Section III. Opinion
Terrestrial
Question
1 . INDICATOR
Chemical
Organism
Population
Community
Combination
No opinion
Other
Inconclusive
Total
Mean number of indi-
cators/respondent
2. ' NUMBER OF PARAMETERS
None
One
More than one
No opinion
Inconclusive
Total
Number
33
50
74
96
80
2
7
1
343
7
1
90
16
14
128
Percent
26
39
58
75
63
2
5
1
2.7
5
1
70
13
11
100
Freshwater
Number
47
46
49
96
82
1
0
4
325
5
2
97
18
9
131
Percent
36
35
37
73
63
1
0
3
2.5
4
1
74
14
7
100
Marine
Number
12
8
14
19
23
1
2
0
79
0
0
21
2
6
29
Percent
41
28
48
66
79
3
6
0
2.7
0
0
72
7
21
100
Undecided
Number
17
15
20
23
26
5
4
1
111
3
0
28
7
7
45
Percent
38
33
44
51
58
11
9
2
2.5
7
0
63
15
15
100
-------
APPENDIX C (continued)
Terrestrial
3.
4.
5.
Question
LIMITED TRAINING
No
Yes
Inconclusive
Total
OTHER INDICES
No
Yes
Inconclusive
Total
RATING
Poor
Fair
Good
Excellent
Inconclusive
Total
Number
64
38
25
127
49
51
28
128
2
22
51
34
19
128
Percent
50
31
19
100
38
40
22
100
1
17
40
27
15
100
Freshwater
Number
75
37
19
131
51
58
22
131
6
30
54
27
14
131
Percent
57
28
15
100
39
44
17
100
4
23
41
21
11
100
Marine
Number
19
6
4
29
6
18
5
29
2
3
10
8
6
29
Percent
65
21
14
100
21
62
17
100
7
10
34
28
21
100
Undecided
Number
27
7
11
45
11
20
14
45
3
9
11
5
17
45
Percent
60
16
24
100
24
45
31
100
7
20
24
11
38
100
-------
APPENDIX D
DATA SUMMARY OF WRITTEN INQUIRIES BY YEARS OF EXPERIENCE
1
n
n
Question
CRITERIA.USED
Chemical
Laboratory toxicity
Field survey
Other
Inconclusive
Total
PARTICIPATION
Never
Occasionally
Frequently
Inconclusive
Total
INDICES USED
Terrestrial
Freshwater
Marine
Total
Percent respondents
using index on one
or more habitats
Section II. Type of involvement
0 - 2 2-5 5-10
Number Percent
2
1
8
4
5
20
10
5
40
20
25
100
6
8
2
4
20
7
8
13
13
30
40
10
20
100
65
Number
3
5
39
5
19
71
12
33
23
3
71
30
31
30
38
Percent
4
7
55
7
27
100
17
47
32
4
100
54
Number
5
1
42
4
26
78
7
47
23
1
78
32
45
16
69
Percent
7
1
54
5
33
100
9
60
30
1
100
88
10+
Number
4
1
89
10
55
159
9
80
57
13
159
94
80
28
146
Percent
2
1
56
6
35
100
6
50
36
8
100
92
Undecided
Number
0
1
1
1
3
6
1
3
1
1
6
2
0
0
2
Percent
0
17
17
17
49
100
17
49
17
17
100
33
1
Responses falling on a division were elevated to the next higher rank.
-------
APPENDIX D (continued)
Question
INDICATOR
Chemical
Organism
Population
Community
Combination
No opinion
Other
Inconclusive
Total
Mean number of,indi-
cators/respondent
NUMBER OF PARAMETERS
None
One
More than one
No opinion
Inconclusive
Total
LIMITED TRAINING
No
Yes
Inconclusive
Total
0-2
Section III. Opinion
2 - 5
5-10
10+
Undecided
11
10
12
15
15
3
0
0
66
55
50
60
75
75
15
0
0
19
15
21
48
31
1
5
5
145
27
21
30
68
44
1
7
7
28
27
37
51
48
4
1
3
199
36
35
47
65
62
5
1
4
46
60
83
112
114
0
6
3
424
29
38
52
70
72
0
4
2
4
5
5
5
5
0
0
0
24
67
83
83
83
83
0
0
0
-------
APPENDIX D (continued)
4.
5.
Question
OTHER INDICES
No
Yes
Inconclusive
Total
RATING
Poor
Fair
Good
Excellent
Inconclusive
Total
0
Number
14
3
3
20
0
3
5
4
8
20
- 2
Percent
70
15
15
100
0
15
25
20
40
100
2
Number
28
32
11
71
2
12
31
19
7
71
- 5
Percent
40
45
15
100
2
17
44
27
10
100
5
Number
21
36
21
78
4
15
28
17
14
78
- 10
Percent
27
46
27
100
5
19
36
22
18
100
10+
Number
51
78
30
159
6
37
61
33
22
159
Percent
32
49
19
100
4
23
38
21
14
100
Undecided
Number
2
1
3
6
0
0
4
0
2
6
Percent
33
17
50
100
0
0
67
0
33
100
------- |