SPECIES DIVERSITY IN STRESSED
MARINE COMMUNITIES
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
Pacific NW Environmental Research Laboratory, NERC-Corvallis
subject: Task Completion: Report on Diversity Indices date:December 17, 1974
D. J. Baumgartner, Chief
FROM: Coastal Pollution Branch
TO: Director, PNERL
The enclosed compendium of reports represents completion of
Milestone No. 1 in Task 2, ROAP 21 BEI. In addition, it is partially
responsive to the second milestone in that task which is due in
March 1975. A subsequent report will be prepared specifically for
that milestone.
Copies of this compendium should be sent to Dr. McErlean and
to the Office of Enforcement and General Council because they represent
the responsible parties who submitted the statement of need for
which this ROAP was designed. In addition, a copy of the report
should be sent to Director, National Ecological Research Laboratory,
because he has responsibility to develop community structure indices
relating to pollution of atmospheric and terrestrial resources.
One of NERL's objectives is to determine how a unifying scheme of
diversity indices may be applied to all ecosystem components.
The individual reports have been offered for publication in
scientific journals or are to be submitted in the near future.
Because of this, this compendium should not be cited or reproduced
for general distribution. Questions about the availability of copies
in the meantime can be referred to Dr. Rick Swartz, to the other
authors or to me. As soon as we have definite information on the
time and place of publication, we will report that to the recipients
of this report and will make a general notification in our research
highlights report.
Only ten copies of this compendium have been prepared.
Attachment
cc: R. Swartz
P. Lefcourt
EPA form 1320-6 (Rev. «-72)
-------
SPECIES DIVERSITY IN STRESSED MARINE COMMUNITIES
A Report Prepared under ROAP 21 BE I,
Biological Indices for Marine Ecosystems
Coastal Pollution Branch
Pacific Northwest Environmental Research Laboratory
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
December 197^
-------
CONTENTS
Application of Species Diversity Patterns in Marine Pollution 1-40
Investigations
Richard C. Swartz
Diversity Indices as Criteria for Biological Responses in 41-59
Stressed Ecosystems
Charles S. Greene
Impact of Pulp Mill Effluents on Estuarine and Coastal Fishes 60-126
Robert J. Livingston
A Review of Clustering Techniques with Emphasis on Benthic 127-148
Ecology
John D. Walker
i
-------
PREFACE
These contributions were supported wholly or in part by the
Environmental Protection Agency's Research Objective Achievement
Plan 2IBEI, "Biological Indices for Marine Ecosystems." The major
objective of this plan is to develop, evaluate, and recommend
biological survey designs and data analysis procedures for relating
the condition and stability of marine communities to environmental
qudIi ty.
This report is concerned primarily with the use of species
diversity indices in pollution research. I examined the different
kinds of indices, ecological significance of the diversity concept,
and examples of the application of diversity patterns in pollution
biology. The correlation between benthic diversity and abiotic
environmental factors in the Southern California shelf and slope was
assessed by Dr. Greene. Dr. Livingston used diversity indices and
other population and community characteristics in his study of the
effects of pulp mill effluents on estuarine and coastal fishes. Mr.
Walker reviewed a variety of clustering techniques for discriminat-
ing different biotic assemblages. All of these contributions will
appear in the scientific literature.
Richard C. Swartz
-------
APPLICATION OF SPECIES DIVERSITY PATTERNS
IN MARINE POLLUTION INVESTIGATIONS
by
Richard C. Swartz
Coastal Pollution Branch
Pacific Northwest Environmental Research Laboratory
Marine Science Center
Newport, Oregon 97365
October 1974
1
-------
INTRODUCTION
Spatial-temporal changes in the number and kinds of species
occur in all ecosystems and have frequently been described in the
scientific literature. Since the publication of G. E. Hutchinson's
(1959) article subtitled "Why are there so many kinds of species?",
fundamental knowledge of the relation between diversity patterns
and causal environmental factors has increased sharply. Concomitant
interest in the use of species diversity as an indicator of bio-
logical conditions in polluted ecosystems led to a plethora of
reports which utilized biomathematical indices of diversity often
with little apparent understanding of their basic or applied signifi-
cance. This review will examine several concepts of diversity,
environmental determinants of diversity, a variety of diversity
indices, and examples of their judicious application in marine pol-
lution studies.
CONCEPTS OF DIVERSITY
The classic concept of diversity is simply the number of species
in an area or collection. This aspect of diversity is often termed
species richness. Diversity is also considered a function of the
distribution of individuals among the species (Lloyd and Ghelardi,
1964). Richness and species frequency distribution are fundamentally
different aspects of community structure. The presence of a species,
no matter how rare, indicates the existence of a unique functional
role. Richness is thus a measure of niche diversity. The species
-2-
-------
frequency distribution defines the extent to which the assemblage is
dominated by the most important species (Whittaker, 1965). When
dominance concentration is high, the "effective" diversity is reduced.
If individuals are randomly selected from assemblages of equal
richness, more species will be found per number of individuals in the
assemblage with lowest dominance concentration.
Richness and dominance concentration are not necessarily cor-
related (Table 1). The distribution of individuals among species can
change markedly without an accompanying increase or decrease in the
number of species. Spatial-temporal patterns of the two components
of diversity are often quite different in marine assemblages [macro-
benthos (Boesch 1972, 1973); eelgrass epifauna (Marsh 1970); demersal
fishes (Swartz, DeBen and McErlean 1974)]. From both theoretical and
empirical observations, richness and dominance concentration should
be analyzed separately and mathematical indices for the two should be
independent of one another.
RICHNESS INDICES
The number of species (S) is dependent on the total number of
individuals collected (N) and the quantity of habitat sampled. S can
be used as a richness index only if it is adjusted for differences in
sampling size. This can be accomplished by assuming a constant rela-
tion between S and N, e.£. logarithmic, or by estimating the number
of species that would have appeared in each sample if all samples were
of the same size (rarefaction). Several indices have been proposed
-3-
-------
Table 1. Relation between richness and dominance concentration
in four hypothetical collections.
Species A
z 10
y 10
x 10
w 10
v 10
u 10
t 10
s 10
r 10
q 10
Total individuals 100
Relative diversity:
Richness high
Dominance
concentration low
Diversity indices:
Richness (s) 10
Dominance
concentration 0.09
(Simpson1s
Index, p. 19)
Number of individuals
per collection
B CD
80 20 82
10 20 8
2 20 5
2 20 3
1 20 1
1
1
1
1
1
100 100 100
high low low
high low high
10 5 5
0.65 0.19 0.68
-------
for each method.
Indices which assume a theoretical S-N relation
(S-l)/ln N
Gleason (1922), Margalef (1960)
S/log N
log S/log N
S/ OF
Menhinick (1964)
Menhinick (1964)
Menhinick (1964)
S = a 1 n[ (N/cc) + 1]
Fisher, Corbet and Williams (1943)
a = diversity index
These richness indices are useful only when the actual species
frequency distribution of all samples to be compared closely fits
one of the theoretical models. Williams (1964) reported that a was
a good measure of diversity in his lepidopteran collections and
Menhinick (1964) found that insect diversity in lespedeza fields
could be described by S/ vfiT. To test the effects of sample size on
index values for a marine assemblage, Hurlburt's equation (p. 9)
was used to predict the mean number of species that would appear in
random samples of different size from a demersal fish and inverte-
brate collection (Table 2). The value of all of the indices was
dependent on sample size (Table 3). Since there is no reason to
believe that observed distributions will necessarily fit any of the
theoretical models, richness is best estimated by one of the rarefac-
tion methods.
Rarefaction
Sanders (1968) presented a method (rarefaction) for estimating
the number of species which would be present in a sample of any size
-5-
-------
Table 2. Collection "I". The sum of 42 bimonthly trawls at the
mouth of the Yaquina River estuary, Oregon.
Species
Enophrys bison
Embiotoca lateralis
Hexagrammos decagrammus
Platichthys stellatus
Parophrys vetulus
Crago nigricauda
Citharichthys sordidus
Cymatogaster aggregata
Psettichthys melanostictus
Cancer magi ster
Rhacochilus vacca
Phanerodon furcatus
Sebastodes melanops
Scorpaeni chthys marmoratus
Cancer productus
Hexagrammos superci1iosus
Spirontocaris paludicola
Crago franci scorum
Leptocottus armatus
Neomysis mercedi s
Pallasina barbata
Apodichthys flavidus
Crago stylirostris
A1losmerus elonqatus
Pleuronichthys coenosus
Syngnathus griseolineatus
Sebastodes caurinus
Pholis ornata
Pandalus platyceros
Spironitocaris brevirostris
Number of Individuals
920
267
181
178
156
117
104
101
79
58
41
35
29
15
13
11
8
6
6
5
4
4
3
3
3
3
2
2
1
1
-6-
-------
Liparis florae 1
Raja rhina 1
Hexaqraminos stelleri 1
Hyperprosopon argenteum 1
Qphiodon elonqatus 1
Spirinchus dilatus 1
Artedius harrinqtoni 1
Alosa sapidissima 1
Microqadus proximus 1
Total Number of Individuals 2365
Total Number of Species 39
-7-
-------
Table 3. Sample size dependence of richness indices which assume a
theoretical S-N relation. SR is the mean number of
species that would appear in random samples (N = 50 to
2000) from Collection I (Table 2). SR's were computed
by Hurlburt's equation (p. 9).
Sample Data L Richness Indices
N
SR
Ct
(S-l)/lnN
S/logN
logS/logN
s/ Jn
50
12.71
5.50
2.99
7.48
0.65
1.80
100
15.86
5.31
3.23
7.93
0.60
1.59
250
20.64
5.34
3.74
8.61
0.55
1.31
500
25.27
5.62
4.07
9.36
0.52
1.13
1000
30.84
6.03
4.46
10.28
0.50
0.98
1500
34.47
6.29
4.58
10.85
0.48
0.89
2000
37.24
6.49
4.77
11.28
0.48
0.83
2365
39.00
6.63
4.89
11.65
0.47
0.80
-8-
-------
smaller than the original. The percent of the total N represented
by each species in the original sample (ni/N) is calculated and
compared with the percent that a single individual would represent
in the reduced sample (Vnr). It is assumed that all species for
which ni/N would be included in the reduced sample. In addi-
n 1
tion, each species for which i/N < NR would contribute a fraction
of a species (ni/N)(N^) to the number of species in the smaller
sample (SR). For comparative purposes, SR can be calculated for the
N of the smallest sample or a plot of SR against NR can be con-
structed for each sample by rarefaction interpolation. A rarefac-
tion example is given in Table 4.
Hurlburt (1971), Fager (1972), and Simberloff (1972) demon-
strated that Sanders' (1968) rarefaction method almost always over-
estimates the number of species that would occur in smaller samples
if individuals were randomly drawn from the original sample. The
reason is that Sanders' method assumes that individuals of the i—
species would be evenly rather than randomly distributed among
smaller samples from the parent population. For example, given
N = 1000, n. = 10, Nd = 100, Sanders' method predicts that one in-
1 K
dividual of the i— species would appear in every 100 individual
sample, and that the contribution of the i— species to SR(SR )
would be 1.0. If 100 individuals were drawn randomly, however,
th
some samples would contain 0 individuals of the i— species, and
the mean SD would be < 1.0, about 0.65.
i
Sanders' method does correctly predict the SR in two
-9-
-------
Table 4. Example of Sanders' (1968) rarefaction method
of predicting the number of species in a
reduced sample.
Total
Species
n.
ni/N
z
174
0.4901
y
98
0.2761
X
37
0.1042
w
19
0.0535
V
12
0.0338
u
8
0.0225
t
3
0.0085
s
2
0.0056
r
1
0.0028
q
1
0.0028
10
355
0.9999
S = 10
N = 355
Nr = 85
1/Nr = 0.0118
SR = 6 + (85)(0.0085 + 0.0056 + 0.0028 + 0.0028) = 7.67
-10-
-------
instances:
if n. = 1; SR = NR/N
if n. > N-Nr; Sr = 1.0
+ h
In the first instance the i— species is rare and in the second it
is so abundant that the reduction in sample size is less than the
t h
original n^; thus the i—species appearance in the smaller sample
is certain. The degree of error in the predicted SR increases
i
between the extremes of very abundant and very rare species. It
follows that the method will be most misleading for those samples
in which most of the species are neither rare nor abundant, i_.e.
for samples with relatively high evenness.
Simberloff (1973) and Fager (1973) recommended that SR be
empirically estimated by many random selections of NR from N, a
procedure used by Menhinick (1964). Hurlburt (1971) gave an
equation for the prediction:
(N-n.)!
Sr • 7 1
s
i = l
NR![(N-n.)-NR]!
Nr!(N-Nr)!
= S ~^_log _1 ^[log (N-ni)!-log(N-ni-NR)!] - [log N!-1 og(N-NR)!]
when n^>N-NR, the contribution to the summation is zero.
when n^l, the contribution is 1-(NR/N).
A table of log X! for X £ 1050 is available in Lloyd, Zar and
Karr (1968).
-11-
-------
The three methods of calculating SR are compared in Table 5.
Murlburt's (1971) equation gives the exact value of mean SR for
random sampling. As noted above, Sanders' (1968) rarefaction
method overestimates this value. Empirical random sampling gave
reasonably accurate SR estimates which would have been improved
if more replicates had been taken.
The SR of a collection can be higher than that of a second
collection at one NR, but lower at another NR (Hurlburt, 1971).
This can be true even when the original number of species and in-
dividuals are the same for both collections (Table 6). The prob-
ability that abundant species will appear in reduced samples
rapidly approaches 1.0 as NR increases, but that probability for
less abundant species continues to increase at even higher NR's.
Thus, changes in relative species richness are dependent on the
original species frequency distributions. Comparisons of SR's
at a single NR can be misleading.
The rarefaction methods assume that the individuals of all
species have a random spatial distribution. In fact, many marine
organisms show clumped or aggregated distributions. Fager (1972)
used a computer simulation of clumping to demonstrate that for 1000
individual samples from an original collection of S=110, N=6664,
the predicted SR decreased from 84 to 39 species as the mean clump
size increased from 1 (random distribution) to 16 individuals
(highly aggregated distribution). Fager also showed that the median
observed S in real benthic macrofaunal samples was less than the
-12-
-------
Table 5. Comparison of the estimated number of species (SR)
in samples smaller than the original population
(Collection I; S = 39, N = 2365) as determined by
Sanders' (1968) rarefaction method, ten random
samples for each NR, and Hurlburt's (1971) equation.
Estimate
Sanders' Method Random Sampling Hurlburt's Equation
50 14.3 12.4 12.7
100 17.2 16.4 15.9
250 22.4 20.9 20.6
500 27.4 25.2 25.3
1000 32.3 30.9 30.8
1500 35.0 34.6 34.5
2000 37.4 35.8 37.2
-13-
-------
Table 6. Influence of reduced sample size on estimates of
relative species richness (Hurlburt's SR) in two
hypothetical collections.
Col lection
Species A B
z 500 200
y 420 200
x 10 200
w 10 200
v 10 195
u 10 1
t 10 1
s 10 1
r 10 1
q 10 1
N 1000 1000
S 10 10
Sr(Nr=10) 2.76 4.51
SR(NR=100) 7.22 5.50
-14-
-------
predicted SR based on a random distribution. Thus, unless one can
adjust for non-random distributions, SR's calculated by Sanders'
method, random multiple sampling, or Hurlburt's equation will over-
estimate the number of species in reduced samples. Further, even
the relative richness of different collections at a common N could
be misleading if the spatial distribution patterns were appreciably
different.
Numerical and Areal Richness
Species richness can be compared for samples of common number
of individuals (NR; numerical richness) or common sampling effort
(area, volume, tow duration; areal richness) (Hurlburt, 1971). In
light of the difficulty of accurately predicting SR for a given N,
areal richness estimates have a considerable practical advantage
in that common sampling effort can be incorporated into survey
designs while it is virtually impossible to obtain samples containing
the same number of individuals. If sampling effort is constant,
observed S's can be directly compared. If effort varies, SR can
be determined for the number of individuals that would have been
present if the sampling effort for each collection had been that
of the smallest effort in the survey (Swartz, DeBen, and McErlean,
1974). Each collection would have a unique NR, calculated on the
assumption that catch of individuals is directly proportional to
effort.
Areal richness is a measure of species density and it is
certainly different from numerical richness. Interpretation of
-15-
-------
diversity patterns can be strongly affected by the choice of the
comparative base for richness estimates, i_.e^ common effort or N.
This is illustrated by temporal diversity trends in a polychaete
assemblage of the New York Bight (Table 7). During the first three
of six quarterly benthic surveys of a site 12 km south of Fire
Island, New York, dominance concentration among the polychaete
species was relatively low (l-Simpson's index = 0.86-0.90). After
the third cruise a previously rare species (Spiophanes bombyx)
became very abundant and dominance concentration increased
(1-S.I. = 0.52-0.68). The appearance of S_. bombyx was accompanied
by little change in areal richness (15-18 species after the first
survey with no temporal trend) but by a precipitous drop in
numerical richness (from > 17 species in the first three surveys
to < 12 species in the last three).
How should the temporal richness pattern in this polychaete
assemblage be interpreted? Again, diversity embraces two different
ecological concepts: dominance concentration and the number of
species. Estimates of each should be independent of the other.
Numerical richness, however, is a function of the distribution of
individuals among the species. Its decline after the appearance of
Spiophanes in large numbers is but an indirect measure of the in-
creased dominance concentration. The estimate of the actual number
of species in the assemblage (areal richness, species density, niche
diversity) is not biased by the species frequency distribution and
it did not change significantly after the third cruise.
-16-
-------
Table 7. Influence of a temporal change in evenness on the areal and numerical richness
of a polychaete assemblage in the New York Bight.
Spiophanes bombyx
Total number of individuals
Number of species
Dominance Cone. (l-Simpson's Index)
SR (N=81) (Numerical richness)
2
Sample area (m )
O
Sp (0.25m ) (Areal richness)
1972 1973 1974
December February May August December February
2
0
15
193
180
161 •
171
81
122
283
487
340
30
19
19
18
18
19
0.89
0.90
0.86
0.52
0.66
0.68
24.0
19.0
17.7
11.7
11.5
10.8
.60
.50
.50
.25
.50
.30
23.2
15.0
17.0
18.0
16.5
17.6
-------
In summary, species catch per unit effort provides the best
estimate of species richness. If sampling effort varies,
Hurlburt's (1971) equation can be used to predict SR for a common
effort, but the possibility that nonrandom spatial patterns will
result in an erroneously high must be considered.
INDICES OF DOMINANCE CONCENTRATION
Determination of the concentration of dominance requires a
measure of the relative importance of individual species. The
number of individuals, biomass (wet, dry, ash-free weights),
caloric content, or some measure of functional significance can be
used as importance criteria. Importance values based on biomass
have the disadvantage that rare, large species occasionally create
extreme values that are difficult to handle quantitatively. How-
ever, biomass is a good measure of the importance of marine plants
and colonial animals which can't be counted easily. Hurlburt (1971)
suggested that the importance of a species could best be defined
as the sum over all species that would occur when a particular
species was removed. This definition has great theoretical appeal
because it would disclose the significance of sometimes rare species
which control diversity and abundance of large assemblages through
predation, primary production, cleaning symbioses, etc. In prac-
tice, analysis of the functional significance of all species in a
community would be very difficult. Most faunal diversity studies
are based on numeric abundance, which should be adequate for
-18-
-------
comparative studies of structural changes in response to pollutional
stress.
The extent to which an assemblage is dominated by the most im-
portant species is determined by its species frequency distribution.
Unfortunately, no single statistical property of that distribution
can be accepted as an unequivocal index of dominance. There are,
however, several indices which are sensitive to changes in dominance
patterns. These will be reviewed with respect to their dependence
on the relative importance of the dominant species and their independ-
ence of sample size or the presence-absence of rare species.
The effects of sample size were examined by randomly drawing
from a parent population (Collection I, Table 2, S=39, N=2365) ten
samples of each of the following sizes, N=50, 100, 250, 500, 1000,
1500, and 2000. The mean value of each index at each sample size
is given in Table 8.
Dominance concentration is not greatly affected by the presence
or absence of rare species. For example, the two distributions
100,100 and 100,100,10,5,3,1,1 are both dominated by two species.
To test the influence of rare species on index values, the 22 least
abundant species (n^<10) in Collection I were excluded and the
indices recalculated (Table 8).
Relative Abundance of Dominant Species (RADS)
Dominance is usually concentrated in the first two or three
most abundant species. An obvious measure of dominance concentra-
tion is thus the relative proportion of all individuals represented
-19-
-------
Table 8. Influence of sample size on diversity indices. Index values for sample sizes 50
through 2000 are the mean of 10 random samples from Collection I (Table 2).
Indices for N = 2365 are the actual values for Collection I. Values at N=2305
are for Collection I excluding the 23 least abundant species (n^lO).
nl
(n1+n2+n3)
1-Simpson's
Standard
Number of
N
S
N
N
Index
Deviation
Moves
H'
H
J'
J
50
12.4
37.2
62.2
0.82
0.54
0.16
0.88
0.75
0.81
0.80
100
16.4
40.0
61.2
0.80
0.53
0.21
0.92
0.82
0.76
0.75
250
20.9
39.3
58.5
0.81
0.58
0.24
0.95
0.90
0.73
0.72
500
25.2
38.6
58.4
0.81
0.59
0.20
0.97
0.93
0.69
0.69
1000
30.9
38.8
58.1
0.81
0.59
0.19
0.98
0.95
0.65
0.65
1500
34.6
39.3
58.2
0.81
0.58
0.18
0.97
0.96
0.63
0.63
2000
35.8
38.9
58.1
0.81
0.59
0.16
0.98
0.96
0.63
0.63
2365
39
38.9
57.9
0.81
0.58
0.16
0.98
0.96
0.61
0.58
2305
16
39.9
59.4
0.80
0.62
0.37
0.92
0.91
0.76
0.76
-------
by the most numerous species. The proportion of the most abundant
nl
(jq—) and three most abundant [(n-| +n2+n2)/N] species in Collection I
was, as expected, essentially independent of sample size, although
these estimates were more variable at smaller sample sizes (Table 8).
Elimination of the 23 least numerous species caused only a slight
increase in RADS values. The number of species to be included in
the RADS estimate should be decided independently for each collection
series.
Simpson's Index
Simpson (1949) suggested that diversity could be expressed as
the probability that two individuals drawn at random and without
replacement from a multispecies assemblage would belong to the
same species. That probability is expressed as Simpson's Index (SI).
For index values to be positively related to diversity they
should be expressed as the complement of Simpson's index (Mcintosh,
1964). Note, however, that 1-SI is inversely related to dominance
concentration. When all species belong to the same species, 1-SI=0;
when they all belong to different species, 1-SI=1.0.
The complement of Simpson's index was essentially independent
of sample size or the presence-absence of rare species (Table 8).
The value of this index for skewed distributions is strongly dependent
on the relative abundance of the first few most abundant species.
SI
-21-
-------
Standard Deviation (SD)
The standard deviation of the mean number of individuals per
species provides a measure of the species frequency distribution
pattern. Fager (1972) suggested that sample size dependence could
be eliminated by scaling SD on the basis of the maximum range for a
given N and S. Thus, scaled SD = (maximum-observed)/maximum, where
the maximum value is (N-S)/ Js". Scaled SD was essentially constant
for samples from Collection I of N>100 (Table 8). The removal of
rare species had little effect on scaled SD, but this might not be
true for all species frequency distributions.
Number of Moves (NM)
Fager (1972) proposed an index based on the number of "moves"
required to change an observed species frequency distribution into
an even distribution. In essence, individuals are "moved" from
the more abundant to less abundant species until all have the same
number of individuals.
NM = N (S+l) - 2- R,- n,
2 i=l
where is the rank of the i— species when all species
are arranged in order of decreasing abundance.
Scaled NM = (maximum-observed)/maximum
maximum NM = (N-S)(S-l)/2
NM showed a rather peculiar dependence on sample size, reaching a
maximum value at N = 250 (Table 8). This index is not a good
-22-
-------
indicator of dominance concentration because it is greatly affected
by presence-absence of rare species.
Diversity Indices from Information Theory
i
The Shannon-Weaver equation (H1) provides a measure of the un-
certainty of species identification of an individual picked at
random from a multi-species assemblage (Pielou, 1969). Since this
uncertainty increases as both richness and the evenness of distribu-
tion of individuals among the species increase, H' is a measure of
"overall" diversity.
S S
H' = -cl !j_ log Jj. = C. (N log N - ^Ln. log n.)
i=l N N N i=l
C is a constant whose value depends on the chosen log base.
For base 10, C = 1.0; for base e, C = 2.302585; for base 2, C =
3.321928 (Lloyd, Zar, and Karr 1968). Obviously, the log base
choice will not alter relative diversities. Log base 10 is used
in this report.
Pielou (1969) noted that the use of H1 is appropriate only
when collections are truly random samples from a parent population
whose boundaries are known. If we do not know that the structure
of an assemblage is homogeneous over a particular space or time,
a collection is not a sample from a larger entity. Such a sample
must be considered a population in itself, and the appropriate
information-theoretic measure of diversity is Brillouin's equation:
S
H = C log N! = C (log N!-^>log n.!)
N" rrpTn^TTTTnjr N i=l
-23-
-------
Despite the possibility that the choice of H or H1 may be
incorrect from a mathematical perspective, the diversity patterns
indicated by both equations are virtually identical. Observed
correlation coefficients between the two exeefed 0.99 (Boesch 1971,
Rex 1973).
The value of H1 and H was not strongly dependent on sample
size, especially for samples of N>250 (Table 8). Both indices were
affected by the exclusion of rare species (n^)<10); H' dropped from
0.98 to 0.93, and H from 0.96 to 0.92 (Table 8). H and H' are more
sensitive to the toal number of species than is Simpson's index,
but their value is primarily a function of dominance concentration.
Fager (1972) suggested that these indices could be made independent
of sample size by scaling according to (observed H1 - minimum H1)/
maximum H' - minimum H. H'max occurs when individuals are evenly
distributed among the observed number of species.
when all n/s = N/<-
n'—ay = 1 (N log N - S ^ log ?) = log S
IlluA J
H' . occurs when S-l species are represented by one individual
mm
and one species has N-S+l individuals.
H'min = I [N 109 N " (N"S+1) 109 (N"S+1)]
Scaled H' for the example in Table 8 were more dependent on sample
size than observed H'. For example, at N = 500, mean H' = 0.97 and
mean scaled H' = 0.66; at N = 2365 H1 = 0.98 and scaled H' = 0.60.
-24-
-------
Evenness
Pielou (1969) recommended that the evenness of the species
frequency distribution could be estimated independently of the
number of species by dividing the observed H by the maximum pos-
sible H for observed S and N.
J' = H ^H'max = H,/1°9 S
J = H/H,
max
H = & log
max N 3
N!
a?) <3
S-r1
+ 1 !!
N
where [^J is the integer part of
wl M
and r = N - S ^ ; i_.e, if is not
an integer, H is calculated assuming
3 max
S-r species have
IS
individuals and
r species have + 1 species
Pielou's evenness indices (J1 and J) are highly correlated and the
choice of the index would not influence the interpretation of the
evenness patterns.
J and J' are more dependent on sample size and the presence of
rare species than either H or H' (Table 8).
Lloyd and Ghelardi (1965) proposed a measure of evenness based
on the ratio of the number of species (S1) predicted by MacArthur's
broken-stick model for observed H' to the number of species (S) in
the sample: m
e = I
-25-
-------
Lloyd and Ghelardi (1964) selected the species frequency distri-
bution of MacArthur's model rather than complete evenness (H )
r x max'
as the comparative base of their equitability index. MacArthur's
model is close to the maximum attainable diversity of a real
biological assemblage while complete evenness would be unlikely to
occur in nature. Boesch (1971) found no significant differences
between evenness patterns determined by £ and J'. Sheldon (1969)
showed that J1 is less dependent on S than is£, and recommended
J' for general use.
hill (1973) showed that several diversity and evenness
indices could be derived from a single equation:
n,\a W1* ... ~
N / V N
if a = 0, Dq = S
i f a = 1, = exp H'
-1
1
TTiT
if a = 2, D,
rS n '
hi
[>1N.
Dq is the total number of species (richness), H' is the natural
logarithm of D-j, and is the reciprocal of Simpson's (1949)
diversity index.
is an estimate of the "effective" number of species in an
assemblage (Hill, 1973). All of the species will be included at a = 0,
only the most abundant at a = 2. Shannon's index (In D1) lies between
these two diversity numbers and provides little additional information.
-26-
-------
Simpson's index is less dependent on sample size than indices of
lower order (a<2).
The evenness of a collection is defined by the shape of a
histogram in which all n.'s are portrayed, usually in decreasing
order (Pielou, 1969). The width of the histogram is determined by
the number of species, and its maximum height by the maximum n^.
If, after adjustment for differences in S and max n^, the shape of
this histogram is the same for two collections, they possess equal
evenness.
Hill (1973) has shown that evenness can be expressed as the
ratio of any two diversity numbers:
Ea,b = Da/0b
When all n^/N are equal (complete evenness), all diversity numbers
are equal to S. In less even species frequency distributions Dq
will still be S, but the effective number of species at higher orders
will be progressively fewer, i_.e. less at than N^.
Pielou's (1966) "evenness" index (J1) is not a true measure of
evenness (Hill, 1973), i_.e. J = H'/lnS = In D,/ln D , but E, =
I U I
VDo-
Evenness as defined by the shape of an adjusted n^-S histogram
or by E . is a concept of little, if any ecological significance,
fli > D
This is evident from the fact that two collections of radically
different richness and dominance concentration can have the same
evenness. This is true for collection pairs such as 5,5,5,5,5,1,1,1,1,1
(N=30,S=10) and 25,5 (N=30,S=2); or, in the extreme, 1000 (N=1000,S=1)
and 1,1,1... 1(N=1000,S=1000).
-------
Summary - Dominance Indices
The effective diversity of a community is a function of its
species frequency distribution pattern. The critical aspect of
that pattern which determines effective diversity is the relative
importance of the highest ranking species. The total number of
species and the evenness of distribution of individuals over all
species have little influence on effective diversity. Thus I have
reviewed diversity indices in terms of their efficacy as indicators
of dominance concentration.
Three indices (Shannon-Weaver's H1, the complement of Simpson's
index, and the relative abundance of the dominant species, RADS)
provide good measures of dominance concentration that are independent
of sample size. Their slight dependence on the presence-absence of
rare species is appropriately related to the small influence of
such species on the total number of individuals in a collection.
For skewed distributions, the value of all three indices is strongly
influenced by the relative dominance of the first highest ranking
species. In more even distributions, the importance of lower ranking
species will assume greater significance. I recommend that both H'
and 1-Simpson's index be calculated routinely. RADS could be used
to ascribe differences in index values to the abundance of specific
species.
-28-
-------
ENVIRONMENTAL CONTROL OF DIVERSITY
The observation that many biological assemblages show lati-
tudinal gradients in their diversity patterns has led to several
hypotheses concerning environmental control of diversity (Pianka,
1966).
1. Time. Diversity will tend to increase in time for both
evolutionary (speciation) and ecological (colonization) reasons.
2. Climatic Stability. In stable climates, species can
evolve highly specialized adaptive mechanisms. More narrowly
defined niches permits the coexistence of a greater diversity of
species.
3. Spatial Heterogeneity. Greater microspatial hetero-
geneity of the environment increases the number of potential
niches.
4. Productivity. As productivity increases, more species
can exist.
5. Competition. As competition between species increases,
niches become more specialized.
6. Predation. By preventing competition between species,
predators prevent any single species from monopolizing a limiting
factor and thus allow more prey species to exist.
Only the first three hypotheses concern purely abiotic factors.
Recognizing that the ultimate control of diversity lies in abiotic
environmental characteristics, Sanders (1968) proposed the stability-
time hypothesis. "Where physiological stresses have been historically
-29-
-------
low, biologically accommodated communities have evolved. Constant
physical conditions allow biological stress (intense competition,
unbalanced predator-prey systems, etc.) to be mediated through bio-
logical interactions. Resulting assemblages are stable, complex,
buffered, and characterized by species with narrow environmental
requirements. As the gradient of physiological stresses increases,
resulting from increased physical fluctuations or by increasingly
unfavorable physical condition regardless of fluctuations, the
nature of the community gradually changes from a predominantly
biologically accommodated to a predominantly physically controlled
community. Physically controlled communities are characterized by
a small number of species with broad environmental requirements.
These species must give adaptive priority to a broad spectrum of
physical fluctuations, thus preventing the development of highly
specialized biological interrelationships. Finally, when the stress
conditions exceed the adaptive abilities of the organisms, an
abiotic condition is reached. The number of species present
diminishes continuously along the gradient."
Sanders (1968) restricted his hypothesis to within habitat com-
parisons of diversity, i.e. to environments with the same spatial
heterogeneity. I doubt that the need for species in physically
controlled communities to adapt to abiotic fluctuations necessarily
prevents them from evolving specialized biological interrelationships.
Rather, these are prevented by the unpredictability of biological
events in physically fluctuating environments.
-30-
-------
Hessler and Sanders (1967) and Sanders (1968, 1969) have
evoked the stability-time hypothesis to explain the increase in
soft bottom marine benthic diversity along the depth gradient from
estuarine to deep sea habitats. Rex (1973) found that gastropod
diversity increased from the continental shelf to the abyssal rise,
but then decreased on the abyssal plain. He attributed this de-
crease to extremely low productivity in the abyss. Dayton and
Hessler (1972) believed that nonselective predation rather than the
stability-time hypothesis explains high benthic diversity in the
deep-sea, a suggestion disputed by Grassle and Sanders (1973). It
is apparent that within the limits of evolutionary or ecological
time and abiotic environmental stability, one or more of Pianka's
(1966) biological hypotheses may explain observed diversities.
Paine (1966), for example, has shown that the presence-absence of
a top predator (starfish) controls the diversity of prey species
in the physically-controlled rocky intertidal zone.
EFFECTS OF POLLUTION ON DIVERSITY
On the basis of the stability-time hypothesis, Slobodkin and
Sanders (1969) predicted that "a perturbation that would have little
effect on physically controlled communities may be catastrophic when
applied to biologically accommodated ones." Boesch (1974) has
applied this concept to the relative sensitivity of biologically and
physically accommodated communities to pollutional stress. The
relatively narrow environmental requirements of continental shelf
-31-
-------
and deep-sea species make these assemblages more sensitive to
human perturbation, despite their high diversity, than estuarine
assemblages. Biologically accommodated communities are stable in
the sense of persistence, physically-accommodated communities are
stable because of their resistence to and resiliency from stress.
The belief that higher diversity assemblages are more resistant
to stress is unfounded, although it has been presented recently in
several coastal zone management reports (Offshore Oil Task Group,
Massachusetts Inst, of Technology, 1973, in Boesch, 1974, and
Environmental Research Consultants, Inc., 1974). Boesch's (1974)
thesis must not be used to justify exposing low diversity (e.£.
estuarine) habitats to additional stress. Rather the alternative
of disposal of wastes in physically stable habitats must be care-
fully scrutinized because of the high sensitivy of their biotic
assemblages to pollution. A case in point is the requirement of
Public Law 92-532 that "In designating recommended ocean dumping
sites, the Administrator shall utilize wherever feasible locations
beyond the edge of the Continental Shelf."
Diversity indices are hardly necessary to evaluate biological
conditions in heavily polluted, sometimes abiotic marine habitats.
Their value in applied ecology lies in the detection of alterations
before they approach irreversible limits and in the evaluation of
the effectiveness (predicted or observed) of pollution abatement.
Several investigations have demonstrated that diversity patterns can
elucidate subtle changes in community structure in response to human
-32-
-------
perturbation.
Watling e^t al_. (1974) studied the diversity of the benthic
macrofauna to evaluate the effects of sludge dumping off the
mouth of Delaware Bay. A diverse and abundant fauna occurred in
and around the dump site. The scaled standard deviation diversity
index showed that dominance concentration was very high to the
south and west of the site. Watling et^ al^. (1974) hypothesized
that particulate organic materials from the disposal operation had
accumulated in these areas and resulted in extremely high densities
of the dominant species, Nucula proxima, a deposit feeder. They
did not conclude that this represented serious environmental
damage.
Boesch (1973) compared biotic conditions in the Elizabeth
River, Virginia, an estuary receiving domestic and industrial
wastes, with the biota of comparable habitats (similar in depth,
sediments and salinity) in the adjacent Hampton Roads. Species
diversity in the Elizabeth River (S = 9-21, = 2.06-3.19) was
substantially less than in Hampton Roads (S = 31-55, = 3.16-4.80).
The benthos of the Elizabeth River was certainly altered, but not
severely disrupted. Boesch (1973) concluded that diversity indices
provide a sensitive indicator of low to moderate pollutional stress.
Bechtel and Copeland (1970) found that fish diversity in
Galveston Bay, Texas, increased with distance from the industrially
polluted Houston Ship Canal. A linear relation existed between H'
and the percent Canal wastewater at most stations in the Bay.
-33-
-------
Copeland and Bechtel (1971) found that the distribution of toxicity,
as measured by the inhibition of growth in the blue-green alga
Coccocochlius elebans, was also correlated with H'. They believed
their model could be used to predict the effects on fish diversity
of a change in the waste loadings of the Canal. Also, the effective-
ness of wastewater treatment could be related to diversity through
the algal bioassay.
McErlean et al_. (1973) studied fish diversity in the Patuxent
River estuary, Maryland, from 1963 to 1967 in relation to the
possible effects of an electric power plant at Chalk Point. There
was little evidence for direct localized effects due to thermal
effluents or other activities of the plant. However, they observed
an alarming five-year downward trend in species richness and
dominance diversity indices which, if continued would increase the
vulnerability of the fish community to stress.
In each of these examples, pollutional stress resulted in a
decrease in species richness or an increase in dominance concentra-
tion. One cannot conclude, however, that diversity will always
decrease in environments altered by man. Fish species richness is
often very high in the vicinity of sewage outfalls and thermal plumes
(Turner, Ebert, and Given, 1966; Grimes and Mountain, 1971). Such
effects are certainly deleterious if the trace contaminants in
fishes near outfalls increases or if a power plant shutdown in
winter results in a lethal temperature change.
Diversity patterns represent only one aspect of community
-34-
-------
structure. Most importantly, they are not sensitive to changes in
species composition. Assemblages without any species in common can
have the same dominance concentration and richness. Spatial -
temporal changes in species composition should be analyzed by
classificatory techniques (Boesch 1973; Swartz, DeBen and McErlean
1974).
The density of the entire community and individual species must
be examined in relation to biotic and abiotic parameters. - Diversity
patterns in all of the pollution investigations cited above were
interpreted in terms of the environmental requirements of individual
species. The concept of the indicator species, despite its limita-
tions, has been applied successfully in marine pollution biology
(Reish 1959, Swartz 1974). Many autecological characteristics (size
frequency distribution, growth and reproductive rates, trophic posi-
tion, pollutant body burden, etc.) can aid in explaining population
response to stress.
Diversity patterns can demonstrate biological alterations in
stressed ecosystems. The correlation of diversity with pollutional
parameters is a much more difficult task. Several investigators
have tried to relate diversity to the toxicity of specific pollutants
or to the gross toxicity of the environment. The validity of these
relationships is often challenged, but they certainly warrant further
study.
Diversity indices do not represent unequivocal criteria of the
status of biological conditons. However, together with knowledge of
-35-
-------
other aspects of community and population dynamics, diversity
patterns can provide a sound ecological basis for regulatory de-
cisions.
ACKNOWLEDGMENTS
I thank Judith Burton for writing the computer programs for
the diversity indices, Douglas Martin for his helpful discussions
about the indices, Faith Cole and Katherine Pinkos for help with
the calculations, and Grace Boden for typing the manuscript.
-36-
-------
REFERENCES FOR
APPLICATION OF SPECIES DIVERSITY PATTERNS
IN MARINE POLLUTION INVESTIGATIONS
-37-
-------
LITERATURE CITED
Bechtel, T. J. and B. J. Copeland. 1970. Fish species diversity
indices as indicators of pollution in Galveston Bay, Texas.
Univ. Texas Contr. Mar. Sci. 15: 103-132.
Boesch, D. F. 1971. Distribution and structure of benthic com-
munities in a gradient estuary. Ph.D. Thesis, College of
William and Mary, Williamsburg, Virginia, 120 p.
. 1972. Species diversity of marine macrobenthos in the
Virginia area. Chesapeake Sci. 13: 206-211.
. 1973. Classification and community structure of macro-
benthos in the Hampton Roads area, Virginia. Mar. Biol. 21:
226-244.
. 1974. Diversity, stability and response to human dis-
turbance in estuarine ecosystems. Proc. First Int'l. Congr.
Ecology, The Hague, Netherlands.
Copeland, B. J. and T. J. Bechtel. 1971. Species diversity and
water quality in Galveston Bay, Texas. Water, Air, and
Soil Pollution 1: 89-105.
Dayton, P. K. and R. R. Hessler. 1972. Role of biological dis-
turbance in maintaining diversity in the deep sea. Deep-sea
Res. 19: 199-208.
Environmental Research Consultants, Inc. 1974. City of Areata:
Challenge of Water Quality Control Plan, North Coastal
Basin IB.
Fager, E. W. 1972. Diversity: a sampling study. Am. Nat. 106:
293-310.
Fisher, R. A., A. S. Corbet, and C. B. Williams. 1943. The
relation between the number of species and the number of in-
dividuals in a random sample of an animal population.
J. Anim. Ecol. 12: 42-58.
Gleason, H. A. 1922. On the relation between species and area.
Ecology 3: 158-162.
Grassle, J. F. and H. L. Sanders. 1973. Life histories and the
role of disturbance. Deep-sea Res. 20: 643-659.
-38-
-------
Grimes, C. B. and J. A. Mountain. 1971. Effects of thermal
effluent upon marine fishes near the Crystal River steam
electric station. Florida Dept. Nat. Resources. Prof. Pap.
Ser. No. 17. 64 p.
Hessler, R. R. and H. L. Sanders. 1967. Faunal diversity in the
deep-sea. Deep-Sea Res* 14: 65-78.
Hill, M. 0. 1973. Diversity and evenness: a unifying notation
and its consequences. Ecology 54: 427-432.
Hurlburt, S. H. 1971. The nonconcept of species diversity: A
critique and alternate parameters. Ecology 52: 577-586.
Hutchinson, G. E. 1959. 'Homage to Santa Rosalia, or why are
there so many kinds of animals. Amer. Natur. 93: 145-159.
Lloyd, M. and R. J. Ghelardi. 1964. A table for calculating the
equitability component of species diversity. J. Anim.
Ecol. 33: 217-225.
, J. H. Zar, and J. R. Karr. 1968. On the calculation of
"information - theoretical measures of diversity. Amer. Midi.
Natur. 79: 257-272.
Margalef, R. 1960. Temporal succession and spatial heterogeneity
in phytoplankton. In_ A. A. Buzzati-Traverso (ed.),
Perspectives in marine biology, Univ. California Press,
Berkeley.
Marsh, G. A. 1970. A seasonal study of Zostera epibiota in the
York River, Virginia. Ph.D. Thesis. College of William and
Mary, Williamsburg, Virginia. 155 p.
McErlean, A. J., S. G. O'Connor, J. A. Mihursky, and C. I. Gibson.
1973. Abundance, diversity and seasonal patterns of
estuarine fish populations. Est. and Coastal Mar. Sci. 1:
19-36.
Menhinick, E. F. 1964. A comparison of some species - individuals
diversity indices applied to samples of field insects.
Ecology 45: 859-861.
Paine, R. T. 1966. Food web complexity and species diversity.
Amer. Natur. 100: 65-75.
Pianka, E. R. 1966. Latitudinal ^riiuients in species diversity:
a review of concepts. Amer. Natur. 100: 33-46.
Pielou, E. C. 1969. An introduction to mathematical ecology.
Wiley-Interscience, New York. 286 p.
-39-
-------
Reish, D. J. 1959. An ecological study of pollution in Los
Angeles - Long Beach Harbors, California. Allan Hancock
Found. Occ. Pap. No. 22. 119 p.
Rex, M. A. 1973. Deep-sea species diversity: decreased gastropod
diversity at abyssal depths. Science 181: 1051-1053.
Sanders, H. L. 1968. Marine benthic diversity: a comparative study.
Amer. Natur. 102: 243-282.
. 1969. Benthic marine diversity and the stability-time
hypothesis. Brookhaven Symp. Biol. No. 22: 71-80.
Sheldon, A. L. 1969. Equitability indices: dependence on the
species count. Ecology 50: 466-467.
Simberloff, D. 1972. Properties of the rarefaction diversity
measurement. Amer. Natur. 106: 414-418.
Simpson, E. H. 1949. Measurement of diversity. Nature 163: 688.
Slobodkin, L. B. and H. L. Sanders. 1969. On the contribution of
environmental predictability to species diversity. Brookhaven
Symp. Biol. No. 22: 82-93.
Swartz, R. C. 1974. Indicator species. In^J. R. Clark (ed.),
Coastal Zone Management Guidebook, Conservation Found.,
Washington, D.C. In press.
, W. A. DeBen, and A. J. McErlean. 1974. Comparison of
species diversity and faunal homogeneity indices as criteria
of change in biological communities. Proc. Conf. Monitoring
the Marine Env., EPA, Washington, D. C. In press.
Turner, C. H., E. E. Ebert, and R. R. Given. 1966. The marine
environment in the vicinity of the Orange County Sanitation
District's ocean outfall. California Fish & Game 52: 28-48.
Watling, L., W. Leathern, P. Kinner, C. Wethe, and D. Maurer. 1974.
Evaluation of sludge dumping off Delaware Bay. Mar. Pol. Bull.5
39-42.
Whittaker, R. H. 1965. Dominance and diversity in land plant
communities. Science 147: 250-260.
Williams, C. B. 1964. Patterns in the balance of nature. Academic
Press, New York. 324 p.
-------
30 September 19 74
Special Task Report
EPA Grant R 801152
DIVERSITY INDICES AS CRITERIA FOR
BIOLOGICAL RESPONSES IN
STRESSED ECOSYSTEMS
Charles S. Greene
Southern California Coastal
Water Research Project
1500 East Imperial Highway
El Segundo, California 90245
(213) 322-3080
-41-
-------
ABSTRACT
Diversity indices are used as measures of the
biological organization of communities. Although
it is assumed that the index values reflect under-
lying environmental factors, these relationships
are seldom established.
Eight diversity indices have been evaluated in
terms of twelve physical/chemical parameters
measured in samples collected from the continen-
tal shelf and slope in an area subjected to
severe man-induced stress. The diversity indices
were found to be highly intercorrelated and, in
many cases, highly correlated with the abiotic
parameters. These indices emphasized three
different components of community organization—
species richness, evenness, and a combination
of these. Stepwise regression analyses showed
that each of these groups of indices was correl-
ated to different sets of the abiotic parameters,
indicating that the different components of com-
munity organization may be responding to differ-
ent aspects of the environment.
Plots of residuals versus the original index
values indicated that the inclusion of additional,
presently undetermined abiotic or biotic variables
would improve the predictive power of the regres-
sion equations.
-42-
-------
INTRODUCTION
The concept of diversity as a meaningful characterization of bio-
logical communities is widely accepted in both theoretical and
applied ecology. Basic to this concept is the idea that within a
biological community, there are relationships between the number
of species, the number of individual organisms, and the number of
organisms per species. Interest in these relationships has prompted
the development of equations for quantifying the diversity of
biological communities.
Diversity indices have been applied to a number of situations, com-
paring changes in community "structure" with respect to time, loca-
tion, and varying degrees of stress. The results of these studies
differ, but in general they support the fundamental concept that
the arrangement of species and organisms within a community is
related to the present or past environmental history of that
community.
The purpose of this study is to examine several commonly used
diversity indices to determine their efficacy as criteria for
detecting changes in community structure resulting from man-induced
stress. Emphasis will be placed on the correlations between
indices of diversity and abiotic (physical/chemical) parameters.
Eight quantitative indices—ranging from the number of species
present to the more complex information theory indices (Shannon
and Weaver 1949; Brillouin 1962)—have been selected for evalua-
tion. Three of these indices (the Shannon-Weaver index, the
Gleason index, and the Simpson index, which are discussed in
the paper) are presently required by State agencies to be
reported in conjunction with data collected during the regu-
lar monitoring of outfall areas in southern California.
-43-
-------
METHODS AND
PROCEDURES
DATA COLLECTION AND ANALYSIS
Biotic Data
Four replicate infauna samples were collected at each of 40 sta-
tions on the continental shelf and slope off the Palos Verdes
Peninsula, Los Angeles County (Figure 1). The samples were col-
lected in August 1973, using a Shipek grab (0.04 sq m) . Biologi-
cal samples were separated using a 1-mm linear aperature sieve.
Each sample was completely analyzed for species and number of
organisms per species by Los Angeles County Sanitation District
biologists. A total of 254 species were identified. Samples
within and between stations were compared for compositional sim-
ilarity, using the so-called "Bray-Curtis" index of dissimilarity,
The results of this comparison showed that, for almost all cases,
the similarity between replicate samples within a station was
much greater than that for replicates between stations. 'There-
fore, replicate samples within each station were pooled.
Physical/Chemical (Abiotic) Data
Single samples for physical/chemical data determinations were col-
lected at each station. These samples were analyzed for (1) sed-
iment particle size fractions, which were combined and used in
this study as percent gravel, percent sand, percent silt, percent
clay, and percent fines according to the Wentworth scale (mean
phi grain size (Wentworth scale with phi notation) and descriptive
sediment coarseness (to be referred to as DSC) according to the
Shepard sediment type description are also used), (2) sulfide poten-
tial (Es~, mV) by the procedure of Kalil (1973) in pore water,
(3) total DDT (ppm, dry weight), (4) organic nitrogen (percent,
dry weight) by the Kjeldahl method, (5) total mercury (ppm, dry
weight), and (6) depth.
ANALYTICAL PROCEDURES
Diversity Indices
A troublesome ambiguity exists in current literature and conver-
sation regarding the names of two popular diversity indices. The
Gleason and Shannon-Weaver indices of diversity (defined below)
-44-
-------
Figure I. Benthic sampling
sites on the Palos Verdes
shelf
PALOS VERDES PENINSULA
33°
45'N
A = 30OM
B = 150M
C = 60 M
D= 30 M
NAUTICAL MILE
DEPTH IN METERS
IOC
/lOBw.,
10A
; VERDES PT.
TRANSECTS
1-10
DEPTHS
From Greene 1974.
-------
are both frequently referred to as the Margelef index of diversity.
Marqelef1s contributions to ecology have gained worldwide recogni-
tion. However, to reduce this persistent ambiguity, it is recom-
mended that these indices be referred to in deference to their
sources, i.e., the Gleason index (Gleason 1922) and the Shannon
or Shannon-Weaver index of diversity (Shannon and Weaver 1949).
In the following definitions, n^ is the number of individuals of
species i, N is the total number of individuals and is equal to
I ni, and S is the total number of species.
The "evenness" of a sample is the distribution of individuals
among species: Maximum evenness occurs when ni is the same for
all species, and minimum evenness occurs when all species but one
are represented by a single individual and the remaining species
is represented by many individuals.
Formulas for calculating the maximum and minimum diversity for
various indices are given by Fager (1972).
1. The number of species per station, species richness (SR) ,
is a simple measure of community structure. Because a
number of researchers have used species richness as a
measure of species diversity (see Hurlbert 1971) , it has
been included in the present study.
2. Gleason's index, D = (S - l)/ln N, takes into account only
the total number cf species and individuals, ignoring the
distribution of individuals among species. This statistic
indicates primarily the species richness component of
diversity and is relatively insensitive to changes in N
when N is large (DeBenedictis 1973). D is also sensitive
to changes in sample size.
3. Simpson's index, SI = I ni (ni - 1)/N(N -1), is the proba-
bility that two individuals selected at random will be the
same species (Simpson 1949). Because the values of this
index go the "wrong" way for a measure of diversity, I will
use 1.0 - SI = SI (Fager 1972).
4. SI scaled,
SI(s) = (SI , - SI . )/(SI - SI . ) ,
w calc min" 1 max min
has been suggested by Fager (1972) to allow comparisons
between samples with different N and S. It appears to be
a measure of the evenness component of diversity, since the
influence of the species richness component is removed by
scaling SI.
5. The standard deviation (SD) is one of the most widely used
statistical measures of dispersion. Applied to a sample
from a community of species, it measures the evenness of
the distribution of individuals among species. I will use
-46-
-------
a scaled version of this measure, and, because it also
goes the wrong way for a measure of diversity,
SD(s) = (SD - SD , ) / (SD - SD . ).
v v max calc / max min
SD(s) ranges from 1.0 for complete evenness to 0.0 (Fager
1972) .
6. The Shannon-Weaver index, H1 = -I In p^, where pi = n^/N,'
is an information theoretical measure of mean diversity
per individual (Margelef 1957) based on Shannon's work
(1949)' on information content in communications theory.
This index is sensitive to changes in the species richness
and the evenness of the distribution of individuals among
the species. Shannon's index is probably the most widely
used of all indices of species diversity except for .
species richness (if SR is considered to be a measure of
diversity). H' is only an estimate of the true diversity
of a community and, for this reason, is often inappropri-
ately used in the strict sense of its definition (Pielou
1969) .
7. The scaled version of the Shannon-Weaver index,
H'(s) = (H' , - H1 . )/(H' - H1 . ),
Ccilc ruin max rnin
is a measure of the evenness component of diversity, since
the effect of the species richness component is removed
by scaling the index from 1.0 for complete evenness to 0.0.
8. The Brillouin index,
H = 1/N{ln (N.'/n1! , n2!, . . . n^)},
is also an information theoretical measure of diversity
formulated by Brillouin (1962). H is considered to be the
true diversity of a collection that has been fully censused
and is considered to be a complete biological entity.
Shannon's index is an estimate of H under these conditions.
H is sensitive to changes in species richness, evenness,
and the number of individuals (N) in the collection,
especially when N is small in relation to S.
These indices constitute three different kinds of diversity
measures, based on the components of diversity that each emphasizes:
(1) species richness indices, D and SR, (2) evenness indices,
SI, SI(s), SD(s), and H'(s), and (3) composite or general indices,
H and H'.
Regression' Analyses
The microenvironmental "determinants" of species diversity were
elucidated by linear regression analyses. The estimation proce-
dure was that of least squares for first-order linear models.
Two approaches to identifying relationships between diversity
-47-
-------
indices (dependent variables) arid abiotic parameters (independent
variables) have been utilized. First, linear relationships
between two variables (expressed as correlation coefficients)
were determined; second, linear stepwise regression analysis was
used to establish the "best" regression equation by incorporating
more than one independent variable into the model. In the
present study, the best regression equation is defined as the
equation, assuming intrinsic linearity of the transformed variables,
produced by the stepwise procedure and comprised of a dependent
variable and one or more independent variables that entered and
remained in the equations with significant F values (p < 0.05).
The interrelationships between abiotic variables and their effect
on reducing the variability in the dependent variable was investi-
gated in a preliminary way, using the stepwise procedure. All
computational procedures, discussed fully by Draper and Smith
(1966), were carried out on an IBM 360/91 computer.
Data Transformation
Preliminary transformation of the original abiotic data were
carried out according to the criteria of Cassie and Michael (1968).
The five sediment fractions were transformed by arc sin/X^j, and
organic nitrogen (% N), DDT, mercury, and sulfide potential (Es)
were transformed by /Xij. Descriptive se'diment coarseness (DSC),
mean phi grain size (X $), and depth were not transformed. Slight
modifications to these transformations were made (as explained
later) for the stepwise regression analyses.
These transformations were applied to the data to at least approx-
imately satisfy the assumption of normality and, consequently,
linearity among the variables.
-48-
-------
RESULTS
CORRELATIONS BETWEEN INDICES
To determine the similarities and differences between the diver-
sity indices, correlation coefficients (r) were calculated for
each possible pair, where all values are significant (p < 0.05).,
These are presented in Table 1. The indices have been arranged
and the matrix partitioned (broken lines) into three groups,
where the highest correlations between indices occur in the sub-
matrices along the diagonal. These groups correspond with the
definitions of the indices presented earlier.
The species richness indices, D and SR, are highly correlated
(0.978) and somewhat unique from the evenness indices. The cor-
relation (0.997) between the information theory indices, H and
H', shows that their responses are almost identical for this set
of data. Both H and H' are highly correlated with the other two
groups, although they exhibit stronger correlations with the
measures of evenness which occur as a highly intercorrelated com-
plex of indices.
Table 1. Diversity versus diversity correlation
matrix, 40 stations.
SR
T
D 1
1
H
T
H' 1
SI
SI(s)
SD (s)
H' (s)
SR
1.000
0.978 |
0.6 89
0.662 j
0. 404
0. 378
0. 380
0. 320
D
1
1.000 1
0.787
0.772 |
0.518
0.473
0.473
0.440
H
1
1
1.000
0.997 1
0.916
0. 884
0.889
0. 889
H1
1
a
1.000 [
0.917
0.878
0.881
0. 895
SI
i
1
1.000
0.987
0.978
0.965
SI (s)
!
1
1
1
1.000
0.994
0.959
SD (S)
1
1
1
i
1.000
0.961
H' (s)
1
1
1
1
- —L
1.000
-49-
-------
CORRELATIONS BETWEEN ABIOTIC VARIABLES
Relationships between the spatial distributions of the abiotic
variables are shown in Table 2 as coefficients of correlation
(r). The variables have been arranged into three groups and the
matrix partitioned accordingly (broken lines). Of the three
rroups, only depth can ba considered to be totally independent
of the influence of the outlalls. The submatrices along the
diagonal show the correlations within qroups, the the submatrices
away from the diagonal reveal the correlations between groups.
While the highest values for r .occur within groups of variables
along the diagonal, there are many highly significant correlations,
positive afid negative, between the groups. The most significant
correlations between groups occur between the finer sediment
grain-size fractions and percent nitrogen, DDT, and sulfide
potential (Es)• The latter variables are known to be outfall-
related, and there is evidence (Greene 1974) that, due to floccu-
lation, higher than expected amounts of clays and fine sediment
particles will accumulate around outfalls.
CORRELATIONS BETWEEN DIVERSITY INDICES
AND ABIOTIC VARIABLES
The simplest and most straightforward test for evaluating the
relationships between these indices of diversity and the abiotic
variables is the correlation coefficient (r). Several additional
or alternative transformations of the raw data proved more satis-
factory for these comparisons; they are (1) ln(Yij) for the
Gleason (D) and species richness (SR) indices, (2) ln(Xij) rather
than arc sin/Xj_j for percent sand for all comparisons except D
and SR, and (3) V^tj for descriptive sediment coarseness (DSC) for
the scaled measures of diversity and the Simpson index (SI). These
additional transformations were indicated by examination of the
plots of residuals from the multivariate analyses to be discussed
later.
The results of these comparisons are presented in Table 3; Table 4
shows the distribution and levels of significance for the values
of r in Table 3. Of these values, 75 percent are significant at
p < 0.05, 57 percent at p < 0.01, and 31 percent at p < 0.001,
and at least 66.7 percent of the correlations for each index are
significant (p < 0.05). No significant correlations occurred
between any of the diversity indices and percent gravel and
sulfide potential (Es).
The Gleason index (D), Shannon index (H'), and Brillouin index (H)
were significantly correlated with more of the abiotic variables
(10 of 12) than the other indices, and the scaled Simpson (SI(s))
index with the least (8 of 12). Of the groups of diversity indices,
the information theory indices (H and H1) have more highly signi-
ficant correlations (p < 0.001) than any other group. The species
richness indices (D and SR), while having slightly fewer highly
-50-
-------
Table 2. Abiotic versus abiotic correlation matrix, 40 stations.
1 2
3
4
5
6
7
8
9
10
11
12
1.
% Sand
1.000 0.095
-0.627
-0.928
-0.761 '
-0.822
0.671
-0.233
-0.765
-0.702
-0.651
-0.205
2.
% Gravel
1.000
0.078
-0.115
-0.096
-0.049
-0.109
0.234
0.089
-0.067
-0.124
0.012
3.
% Clay
1.000
0. 733
0. 314
0. 887
-0.496
0.150
0.556
0. 579
0.411
0. 068
4.
% Fines
1.000
0.848
0.919
-0.711
0. 373
0.680
0.606
0.502
0. 033
5.
% Silt
1.000
0.643
-0.684
0.532
0.534
0.352
0. 320
-0.035
6.
X *
1.000
-0.696
0.2*99
0 j 637
0.593
0. 404
0.056
7.
DSC
1.000
-0.427.
-0.661.*
-0.534
-0.399
-0.131
8.
Depth
1.000
0. 021
-0.155
' -0.180
-0.421
9.
% N
i.opo
0.781
0. 810
0. 495
10.
Hg
1.000
0.779
0.537
11.
DDT
1.000
0.623
12.
ES
1.000
i
CJI
-------
Table 3. Diversity versus abiotics correlation
matrix, 40 stations.
-
D
SR
H
H'
SI
SI (s)
SD(s)
H1 (s)
% Sand
% Gravel
% Clay
% Fines
% Silt
X (j)
DSC
Depth _
% Nitrogen
Mercury
DDT
Eg
0.605
-0.2 30
-0.404
-0.580
-0.593
-0.511
0.700
-0.700
-0.513
-0.418
-0.361
-0.134
'0.514 1 0.616
-0.231 -0.183
-0.284 | -0.486
-0.491 -0.591
-0.570 1 -0.515
-0.416 1 -0.553
0.660 1 0.632
-0.767 L_-0.508
-0.383 -0.563
-0.262 | -0.524
-0.205 -0.449
-0.068 1 -0.140
0.619 . 0.512
-0.185 | -0.185
-0.503 -0.390
-0.591 1 -0.433
-0.502 1 -0.336
-0.560 -0.411
0.6221 0.468
-0.4791 -0.244
-0.577 i -0.501
-0.544 -0.512
-0.473 1 -0.415
-0.153 | -0.142
0.463
-0.165
-0.335
-0.385
-0.310
-0.367
0. 440
-0.249
-0.443
-0.448
-0.339
-0.Ill
0.469
-0.154
-0.335
-0.394
-0.320
-0.379
0. 453
-0.226
-0.458
-0.454
-0.336
-0.~112
0. 476
-0.086
-0.384
-0.413
-0.312
-0.401
0. 400
-0.223"
-0.453
-0.486
-0.419
-0.163
Table 4.
Levels of significant correlations for
the diversity versus abiotics correlation
matrix. * < 0.05, ** < 0.01, and *** < 0.001.
D
SR
H
H1
sr
SI (s)
SD (s)
H' (s)
% Sand
% Gravel
% Clay
% Fines
% Silt
x
-------
significant correlations, show the highest individual correlations,
i.e., descriptive sediment coarseness (DSC) and depth. The
evenness indices (SI, H'(s), SD(s), and SI (s)) show only moderately
high and fewer significant values of r.
STEPWISE REGRESSION ANALYSES
The large number of significant correlations found between the
diversity indices and abiotic parameters, plus the high levels of
correlation among the abiotic parameters themselves, suggested
the application of a more complex linear model. Thus, stepwise
regression analysis was employed.
Two sets of data were used for regression with each diversity
index. Data Set I was comprised of the 12 abiotic variables.
Data Set II was comprised of the 12 abiotic variables and the
cross-products for each pair of abiotic variables. The second
set of data was included to provide insight into the interactions
of the independent variables and the effect of these interactions
on the dependent variable.
The equations obtained from regression on Data Set I are presented
in Table 5, and for-Data Set II,- the interaction terms, in Table 6.
All variables included in the equations were entered and were
retained with significant F values (p < 0.05). While these equa-
tions represent the "best" regression equations, the highly signi-
ficant correlations between many of the abiotic variables indicate
that substitutions could be made without seriously altering the
results.
The degree of accuracy for predicting the diversity indices from
these abiotic variables is shown in Table 7, where R2 measures the
proportion of total variability about the mean diversity explained
by regression. The larger R2 is, the better the fitted equation
explains the variation in the data (Draper and Smith 1966). Although
caution must be exercised in applying this statistic, its interpreta-
tion here appears valid.
-53-
-------
Table 5. "Best" regression equations resulting from stepwise
regression analyses of Data Set I.
In D = 2.868 - 1.038(depth) - 0.328(Hg) - 0.027(ES) + £
In SR = 4.264 - 1.052(depth) - 0.036(ES) + 0.062(DSC) +
H = 3.058 - 0.605 (Hg) - 0. 877 (depth) + £
H* = 3.242 - 0 .6 6 4 (Hg) - 0. 897 (depth) + £
HMs) = 0.671 - 0.111 (Hg) - 0.094 (depth) + £
SI = 0.4 52 + 0.083(% sand) + £
SdTs) = 0.254 + 0.069(% sand) + £
SlTs) = 0.458 + 0.076(% sand) + £
R
-
0.908;
R2
-
0. 825
R
=
0.907;
R2
=
0 . 822
R
=
0.794;
R2
--
0. 630
R
=
0.788;
R2
=
0.620
R
=
0.573;
R2
—
0. 328
R
=
0.512 ;
R2
=
0.262
R
=
0. 469 ;
R2
=
0.220
R
=
0.463;
R2
=
0.214
Table 6. "Best" regression equations resulting from stepwise
regression analyses of Data Set II.
In D = 2.477 - 1.162 (% silt x depth) - 0.491(% clay x Hg)
-0. 854 (% gravel x Es) + 3.539 (% gravel) + £
R
_
0.921;
R2
0.848
In SR = 4.157 - 0.019(Hg x Es) + 0.081(DDT x depth)
+ 0.014 (DSC x x <|>) - 1. 657 (% N x depth)
- 0.599 (depth) + £
R
=
0.926;
R2
0. 858
H' = 2. 710 - 0.075 (% clay x DDT) - 0.520 (Hg x depth)
- 0.040 (Ec; x depth) + £
R
0.816;
R2
0. 666
H = 2.553 - 0.065 (% clay x DDT) - 0.491(Hg x depth)
- 0.040(Es x depth) + £
R
=
0.812;
R2
=
0.659
s'lTs) = 0.795 - 0.233 (% fines x Hg) - 0.037 (ES x depth)
+ 0.02 3(% fines x Es) + 0.266 (Hg x depth) + £
R
0.694;
R2
0.482
SI = 0. 883 - 0.100 (% fines x Hg) - 0.732 (% gravel x depth) + £
R
=
0.629 ;
R2
=
0. 396
H'(s) = 0.602 - 0.059(% fines x Hg) - 0.009(Eg x depth) + £
R
=
0.605;
R2
—
0. 367
SD(s) = 0.616 - 0.067(% fines x Hg) - 0.193(% N x depth) + £
R
=
0.587;
R2
=
0. 345
-------
Table 7. Comparison of correlation coefficients (r), multiple
correlation coefficients (R), and squared multiple
correlation coefficients (R2) for the simple linear
and stepwise regression analyses. The abiotic variable
exhibiting the highest correlation with diversity in
the two variable analyses is shown.
D
SR
„
H'
H' (s)
SI
SD (s)
SI (s)
Simple linear
regression
M
(Xj)
0. 700
DSC/
depth
0. 767
depth
0. 632
DSC
0.622
DSC
0. 486
Hg
0. 512
% sand/
Hg
0. 454
% sand
0.448
% sand
Stepwise
regression,
Data Set I
(R)
(R2)
0. 908
0. 825
0. 907
0. 822
0. 794
0.630
0. 788
0.6 30
0.573
0. 328
0. 512
0 . 262
0.469
0.220
0.463
0.214
Stepwise
regression,
Data Set II
(R)
(R2)
0. 921
0. 848
0.926
0. 858
0. 812
0.659
0. 816
0 . 666
0.605
0. 367
0. 629
0. 396
0. 587
0. 345
0. 694
0.482
-------
DISCUSSION
The biotic/abiotic interrelationships that establish the structure
and organization of ecosystems are exceedingly complex. No single
index or group of related indices have evolved to meaningfully
classify this complexity. Various indices of diversity have been
used to evaluate the compositional (species, biomass, or other
measure of importance) organization of communities. Although
these indices can measure different components of a community or
ecosystem, they must not be used to interpret more than they can
actually define. Specifically, the indices measure the influence
of the different components of organization on the overall organi-
zation of the community. They are always insensitive to the iden-
tity of the entities (measures of importance) contributing to this
organization, and, when the overall organization remains constant,
they are insensitive to shifts in dominance among these entities.
Diversity indices are commonly used to measure the organization of
faunal or floral communities. The numerical value of an index is
interpreted against a background of experience or available biolog-
ical data and used to evaluate the status of the community. How-
ever, little information exists for interpreting these biological
indices in terms of specific physical/chemical (abiotic) parameters
interacting with and possibly determining the organization of these
communities.
An important question relative to environmental monitoring has
evolved from considerations similar to the above: Can diversity
indices be satisfactorily employed as criteria for evaluating
changes in the organization of communities resulting from man's
activities? It was not the purpose of this study to answer this
specific question but to provide information about how diversity
indices are related to the physical/chemical parameters of a
highly stressed marine environment. Specifically, I have used
data collected from an area where major changes in community
composition are well documented (Southern California Coastal Water
Research Project 1973, and unpublished data) and related to outfall
effects.
The eight indices considered can be divided into three groups, each
measuring different components of community organization, i.e.,
species richness, evenness, and a combination of these- Correla-
tions between all possible pairs of indices showed that they were
all significantly correlated to one another (p < 0.05) , indicating
that they all provided satisfactory measures of diversity for this
set of infaunal data.
-56-
-------
A high degree of correlation was also found to exist within and
between groups of abiotic parameters (depth, sediment parameters,
and outfall-related parameters). With the exception of depth,
which is independent of the outfalls, the abiotic parameters
reached their highest or lowest values in the vicinity of the
outfalls. Their overall spatial distributions also show the
influence of depth and currents. These general patterns also
apply to the indices of diversity.
Relationships between the diversity indices and abiotic parameters
were evaluated through simple linear regression and correlation
analysis for each pair of variables (diversity and abiotic factors).
The results showed that 75 percent of the comparisons were signi-
ficant (p < 0.05). These results lend support to the general
concept that the composition and organization of communities become
progressively more under the control of physical/chemical factors
as the environment becomes increasing more rigorous. If this con-
cept is applicable to areas of man-induced stress, then one would
expect to find correlations between biotic diversity and causitive
or causitive-related factors in unnaturally stressed areas. This
idea has been touched upon in other studies (e.g., Storrs, Pearson,
and Selleck 1969), but has not been subjected to intensive investi-
gation .
Because of the large number of significant correlations between the
diversity indices and abiotic parameters, stepwise regression anal-
ysis was used to elucidate the relationships between the combined
effects of these abiotic parameters and each diversity index. The
resulting equations explained a greater proportion of the variabil-
ity in the diversity indices and enabled me to more readily iden-
tify the abiotic variables that were influencing each of the
indices. Only the regression analyses based on the first set of
data will be considered in the following discussion. The analyses
based on the interaction terms were included primarily for their
heuristic value.
A measure of the fit of a regression equation is the percentkge
of variation in the dependent variable explained by the regression.
The largest percentages of variation in any of the diversity indices
explained by the stepwise procedure was for the species richness
indices, D and SR. Thus, for this set of data, these indices
would appear to be the best measures of species diversity in terms
of the abiotic environment. The fact that, for D and SR, depth
is the most important independent variable (and that outfall-
related variables are only of secondary importance) sheds some
doubt on the value of these indices as indicators of man-induced
physical/chemical conditions underlying the faunal diversity for
the Palos Verdes outfall area.
The information theory indices, H and H', proved next best in
terms of variability explained by the stepwise regression equations.
As would be expected, both of these indices respond to the same
abiotic variables. In this case, mercury was the most important
-57-
-------
vaiiabJe, and depth the second most important. It would appear
that these indices may be better indicators of underlying environ-
mental factors than the species richness indices.
Of the four indices designated as measures of evenness, the three
scaled indices are definitely measures of evenness, and the high
correlations between these indices and the Simpson index (SI)
indicates that it is also a measure of evenness, at least for this
set of data. As a group, these indices showed lower (but still
significant) correlations with the individual abiotic variables.
Little improvement in these relationships was added by applica-
tion of the stepwise procedure on the first set of data.
The fact that depth is not significantly correlated with the even-
ness indices suggests that they may provide insight into how all
of these indices would perform in an area where depth was not a
major environmental factor, as it is with several major outfalls
in southern California.
Examinations of the plots of residuals versus the dependent varia-
bles indicated that, in all cases, the higher values of diversity
were underestimated and the lowest values overestimated. This
situation was only slight for the species richness indices but
became progressively more apparent for the information indices and
evenness indices. This weakness m the regression equation can
usually be rectified by the inclusion of additional variables
(Draper and Smith 1966). There is no evidence in the present
study to indicate whether or not the variables to be sought and
included are physical/chemical or biological. However, it is
apparent that to solidly establish these procedures, a larger
number of variables must be examined for a variety of situations.
Because the present study was conducted on data collected from an
area known to be subjected to highly stressful conditions, these
findings may be limited in their general application. However
diversity indices have been shown to be significantly correlated
with a number of possible causative environmental factors.
In conclusion, a criticism of the present study and a recommenda-
tion for future studies is warranted. In most cases, a single
abiotic sample was collected at each station and, consequently,
the four replicate biological samples were pooled. It is reason-
able to assume that, had abiotic samples been obtained from each
of the biological samples, the results would have been improved.
J $8-
-------
REFERENCES
Bnllouin, L. 1962 . Science and information theory. New York:
Academic Press, Inc.
Cassie, R. M., and A. D. Michael. 1968. Fauna and sediments of
an mtertidal mudflat: A multivariate analysis. J. Exp. Mar
Biol. Ecol. 2:1-23. ~
DeBenedicts, P. A. 1973. On the correlations between certain
diversity indices. Amer. Natur. 107:295-302.
Draper, N. R., and H. Smith. 1966. Applied regression analvsis
New York! John Wiley.
Fager, E. W. 1972. Diversity: A sampling study. Amer. Natur.
106:293-310.
Gleason, H. A. 1922. On the relation between species and area.
Ecology 3:158-62.
Greene, C. S. 1974. Grain sizes of Palos Verdes sediments. In
1974 Annual Report, Southern California Coastal Water Research
Project, El Segundo, Calif.
Hurlbert, S. H. 19 71. The nonconcept of species diversity: A
critique and alternative parameters. Ecology 52:577-86.
Kali 1, E. K. 1973. Rapid pore water analysis for sediments adja-
cent to reactor discharges. Paper presented at the IAEA symposium
on environmental surveillance around nuclear installations,
Warsaw, Poland, IAEA-SM-180/37.
Margelef, D. R. 1957. Information theory in ecology. Gen. Svst.
3:36-71. 1
Pielou, E. C. 1969. An introduction to mathematical ecology.
New York: John Wiley.
Shannon, C. E., and W. Weaver. 1949. The mathematical theory
of communication. Urbana, 111.: Univ. Illinois Press.
Simpson, E. H. 1949. Measurement of diversity. Nature 163:688.
Storrs, P. N., E. A. Pearson, and R. E. Selleck. 1969. Final
Report, A Comprehensive Study of San Francisco Bay, vol. 6: Water
and sediment quality and waste discharge relationships. SERL
Kept. No. 67-4, Sanitary Engineering Research Laboratory, Univ.
of Calif., Berkeley.
-59-
-------
IMPACT OF PULP MILL EFFLUENTS ON ESTUARINE AND COASTAL FISHES
(APALACHEE BAY, FLORIDA)
Robert J. Livingston
Department of Biological Science
Florida State University
Tallahassee, Florida, USA
-60-
-------
Abstract
A 2 year field study was carried out to determine the impact of kraft
pulp mill effluents on the fish fauna of a shallow bay system in north Florida
(Apalachee Bay, U. S. A.). Offshore areas that received pulp mill effluents
(PME) had significant increases of color and turbidity and reductions in
(benthic) dissolved oxygen when compared to a nearby control area. Estuarine
and marsh fish assemblages in areas of acute impact were severely reduced in
terms of numbers of individuals (N) and species (S). Offshore areas exposed
to varying (chronic) levels of PME were characterized by complex interactions
that included seasonal variations of impact. A broad offshore area showed
reductions in numbers of individuals and species taken per month. However,
the cumulative (annual) number of species taken was the same for polluted and
unpolluted (control) areas due to a recruitement of relatively rare species
in the areas of impact. Such polluted areas showed decreased dominance as well
as qualitative differences in species composition when compared to control areas.
Inshore bay stations that were most severely affected by PME were dominated
by the bay anchovy (Anchoa mitchilli). While species richness and species diver-
sity were lower at the highly stressed stations, in other outlying areas of
moderate impact (reduced N and S) there were no reductions of such parameters
when compared to control areas. Thus, species diversity was not an indicator
of pollution per se, and was useful only when taken in conjunction with various
other parameters. Transition areas (between polluted and unpolluted portions
of the bay) showed substantial (though periodic) increases in N, S, and species
diversity. Equitability indices were unchanged in polluted portions of the Bay.
In general, the effects of PME on offshore fish assemblages appeared to be due
to a complex combination of habitat alteration, reduced benthic productivity,
and individual behavioral reactions. The alterations of fish assemblages were
compared to other studies in this area on benthic macrophytes and invertebrates
in an effort to assess the usefulness of various indices in studies on the long-
term effects of pollution on estuarine and coastal systems.
-61-
-------
Acknowledgement
The author is most appreciative of the effort made by various students;
this is especially true of the work of Theresa Ann Hooks, Michael S.
Zimmerman, and Kenneth L. Heck, Jr. This research was sponsored in part
by the Coastal Coordinating Council (Florida Department of Natural Resources)
and thanks are due to Mr. Bruce Johnson for his help. Additional support
was provided by the Edward Ball Marine Laboratory (Florida State University)
and much is owed to Dr. Robert C. Harriss and his staff. Dr. Ralph C.
Yerger was also a source of help on various phases of this project. The author
also recognized Dr. Gerald van Belle for his aid in the statistical analysis.
Thank3 are due to Dr. Daniel Simberloff for his thoughtful comments and Dr.
Richard C. Swartz (Environmental Protection Agency) for his encouragement and
financial support (EPA Program Element //l BA025). This is contribution num-
ber 28 of the Tallahassee, Sopchoppy, and Gulf Coast Marine Biological
Association.
-62-
-------
INTRODUCTION
Recently, various indices have been used to evaluate the effects of pollution
on assemblages of estuarinc and coastal organisms (McErlean et al. , 1973; Boesch,
1973; Holland ejt aj.. , 1973). Pulp mill effluents (PME) represent a relatively
common form of contamination, especially in the southeastern United States. Such
effluents are composed of various constituents (organic acids, sulfides, sulfites,
lignin, pectin, polythionates, etc.) that often cause significant changes in water
quality (lowered dissolved oxygen, high biochemical oxygen demand, increased color
and turbidity). Due to the variable nature of PME from one mill to the next, and
the extreme complexity of estuarine and coastal ecosystems, few generalizations
can be made concerning the environmental impact of pulp effluents on such systems.
Considerable work has been directed at the effects of PME on aquatic organisms.
This includes general information (Waldichuk, 1962; Courtright and Bond, 1969) as
well as data on specific components of such effluents (Cole, 1935; Van Horn et al.,
1949; Smith et al., 1965; Servizi et al., 1966; Das et al., 1969). The effects of
PME on fish eggs and fry (Smith and Kramer, 1963; Colby and Smith, 1967) and adult
fishes (Fujiya, 1961, 1965; Walden £t al., 1970) have been described. Various
papers have concentrated on commercially important organisms such as salmon (Jones
et al. , 1956; Alderdice and Brett, 1957; Holland et al., 1960; Grande, 1964;
Servizi et al., 1966; Befits and Wilson, 1966; Sprague and McLeese, 1968a, 1968b;
Sprague and Drury, 1969; Das e£ al^, 1969; Ziebel et al., 1970; Hicks and DeWitt,
1971; Shumway and Chadwick, 1971; Webb and Brett, 1972; Wilson, 1972), oysters
(Hopkins et al., 1931; Odlaug, 1948; Woelke, 1960, 1962, 1968), and lobsters
(Sprague and McLeese, 1968a; 1968b). Thus, there is a relatively good literature
on the impact of pulp wastes on individual components of various aquatic systems.
Although occasional field observations have been made, there is little
information on the long-term effects of PME on aquatic communities. The bulk
-63-
-------
of this work has been done in fresh water areas* Wilson (1953) found that kraft
mill effluents (KME) caused increased fungus growth and a general decline of the
benthic community of the Clearwater River (Idaho, U.S.A.). Wiley e£ al. (1957)
showed delines in benthic species due to pulp mill contamination of the lower
Fox River (Wisconsin, U.S.A.). Ito and Kuwada (1964) found decreased standing
crop and species diversity of freshwater invertebrates as a function of distance
from the pulp mill effluents. Goto and Goto (1972) found long term reductions of
fishes in a Japanese river that were attributable to effluents from pulp mills.
The effects of pulp mills on estuarine and marine assemblages have been ana-
lyzed in several studies. Copeland (1966), working in St. Joseph Bay (Florida,
U.S.A.), found localized decreases in zooplankton species diversity and chloro-
phyll A near a pulp mill outfall. Wildish et al. (1972) found little damage
to the sublittoral benthic fauna of the L'Etang Inlet (New Brunswick, Canada)
due to effluents from a neutral sulphite pulp mill. Pearson (1971), after a
5 year study on the macrobenthic fauna of a sea loch system off west Scotland,
noted little effect of PME on biomass or numbers of species. Peer (1972)
noted quantitative changes in the composition of a marine benthic community
due to discharges of PME. Overall, the data would indicate that such
effluents can seriously disrupt fresh water areas such as rivers and streams;
however, aside from certain localized effects, the impact of PME on marine
systems appears to be minimal in the few studies that have been made.
RecenLly, species diversity has been commonly used as an index in
community analysis. Information theory as introduced by Margalef (1958)
and developed by Hairston (1959), MacArthur (1964), Lloyd (1964), Shannon
and Weaver (1963), and Brillouin (1962) has been important in the develop-
ment of this index. Species
-64-
-------
diversity as a function of community structure has been qualified (Pielou,
1966a; 1966b; 1966c) and criticized (Hurlbert, 1971), but is still frequently
used. This is clearly the case in various impact studies; species diversity
has been used with groups such as diatoms (Patrick, 1963), invertebrates
(Reish, 1960; Odum e_t al. , 1963; Ito and Kuwada, 1964; Copeland, 1966; Wilhm
and Dorris, 1966; 1968; O'Connor, 1972; Boesch, 1972), and fishes (Katz and
Caufin, 1953; Tsai, 1968; Armstrong et al., 1970; Bechtel and Copeland, 1970;
McErlean jil., 1973). In most cases, lowered species diversity has been
associated with the stress of various forms of aquatic pollution. However,
with few exceptions (Dahlberg and Odum, 1970; Bechtel and Copeland, 1970;
McErlean £t al^. , 1973), seasonal changes in diversity have not been considered.
Diversity is often related to factors such as stability, energy flow, population
dynamics, dominance, competition, and structural complexity. As yet, the
relationships of these factors in estuarine and coastal systems are not well
understood, and the effects of different forms of pollution on such parameters
remain largely unexplained. The role of species diversity in the assessment of
marine pollution would thus be open to question.
The present study is a comparison of the fish fauna of an unpolluted
drainage system (Econfina River) and a polluted one (Fenholloway River), with
particular emphasis on certain shallow portions of Apalachee Bay (Florida, U.S.A.).
The study area is shown in Fig. 1. Both rivers drain the same swamp area (San
Pedro Bay), and have comparable discharge rates (Fenholloway-125 CFS; Econfina-
134 CFS). Climatic, sedimentary, and drainage features are similar although the
consolidated oyster reefs of the Fenholloway system differ from the scattered
oyster clusters of the Econfina system. Extensive beds of benthic macrophytes
are found in the offshore areas of the Econfina compared to the mud flats and
scattered (sparse) assemblages of benthic plants in the Fenholloway region
-65-
-------
(Livingston et al., 1972; Zimmerman, 1974). From 1954 to 1974, a kraft paper-pulp
mill (the Buckeye Cellulose Corporation) has discharged -200-220 million liters of KME
per day into the Fenholloway River. A river survey just prior to effluent release
(Beck, 1954) revealed high levels of CC>2 and dissolved (ferrous) iron and a healthy
fauna (invertebrates and fishes). Saville (1966) compared the physical, chemical,
and biological components of the Fenholloway system with the unpolluted Wacasassa
system. He found low dissolved oxygen, high turbidity and color, and increased
levels of lignin and total phosphorous in the Fenholloway drainage. Quammen et al.,
(1971), using ATP analysis, found that the pulp wastes discharged into Apalachee Bay
were confined to inshore areas within 1.7 km of the river mouth. It was thought that
offshore oyster bars localized the pulp wastes inshore and southeast of the river.
This study was designed to determine the impact of pulp mill effluents on the
fish fauna of the Fenholloway drainage system by utilizing various indices of
community structure.
MATERIAL AND METHODS
Collection techniques
Fixed sampling stations were established in the rivers, marshes, and offshore
areas of both systems (Fig. 1). Offshore stations, identified by permanent markers
and/or compass sightings on nearby landmarks, were established at comparable positions
on each system. A long-shore transect was also established for instantaneous com-
parison of the different drainage systems. Water samples (surface and bottom) were
taken with a 1 liter Kemmerer bottle. Temperature (°C) was measured with a stick
thermometer and salinity was determined with a temperature-compensated refracto-
meter calibrated periodically with standard sea water. Dissolved oxygen was measured
by the modified Winkler method (Strickland and Parsons, 1968) and oxygen probes
(Portable hydrolab II-A and Y.S.I, model 54 oxygen analyzer). The probes were
calibrated by Winkler determinations. One 24 hour survey of these parameters was
performed. Benthic samples were taken simultaneously at offshore stations of both
-66-
-------
arens. Such stations were monitored at 4 hour intervals. Turbidity was determined
with a battery operated (Hach) water test kit (Formazin standard method) and a Hach
Model 2100 A turbidimeter. Both instruments were routinely cross-calibrated. Color
was measured with a (Hach) APHA platinum-cobalt standard test. Light penetration
was determined with a standard Secchi disk.
River (estuarine) and marsh stations were sampled biologically using various
techniques. Some stations (Ej_, E^, F1# F2, F5b) were blocked off with seines and
poisoned with rotenone. Dead and dying fishes were taken with seines and dipnets.
Sampling at other stations was carried out with dipnets, gill and trammel nets, and
16 foot otter trawls (3/4" mesh wing and body; 1/4" mesh liner). All offshore
stations were sampled with otter trawls: repetitive 2-minute tows were taken at each
station at speeds of 2-3 knots. Collections were randomly taken at comparable times
during the day at any given station. All animals were preserved in 107. formalin,
sorted, identified, and measured (standard length). Collections were made from
April, 1971 to June, 1973 (bi-monthly collections for the first 6 months and monthly
collections thereafter) according to the following schedule:
April, 1971 - June, 1972
10 trawls/station (E7, Eg, F9, FjoJ = inner stations)
2 trawls/station (Eg, E^q, E^, Ej^, Ei3> F^, Fj^; = outer stations)
June, 1972 - May, 1973 (No outer Fenholloway stations taken in May, 1973)
7 trawls/station (inner and outer stations)
July, 1972; November, 1972; April, 1973
7 trawls/station (Ejq, Eg, E13, "^21* "^20' ^11' ^10' ^15' ^17' l^19' *^22' ^23' =
transect stations)
Each station in the Fenholloway offshore area was placed in a comparable position to
its counterpart in the Econfina system with respect to angle and distance from the
the river mouth. In addition to regular sampling sites, collections were made at outer
-67-
-------
stations using hook and line. Larger predators and game fishes not sampled by
otter trawls were also followed by examination of catches made by sports fishermen
at a local (Econfina) fish camp and commercial fishing interests in the area.
Statistical analyses
The repetitive otter trawl samples taken on all offshore stations were analyzed
In various ways. The trawling technique was examined [Livingston et al_., 1972)
using a modification (Simberloff, 1972) of the "rarefaction" analysis (Sanders, 1968)
for the calculation of the expected number of species (ENS) for each trawl tow.
Computations were made at each station on a scaled basis (using the lowest common
number of individuals taken in a series of collections on all stations) and an
unsealed one (using the actual numbers of individuals taken). Stations where less
than 20 individuals were taken were not analyzed. The following formulas were used
to compute Brillouin (H) and Shannon-Weaver (H') diversity indices (Brillouin, 1962;
Shannon and Weaver, 1963).
Two evenness factors were also calculated: J(Pielou, 1966) and J' (from H').
S
H a (l/N ) (Log NJ - £ Log Hi I)
i= 1
S
s
l/N (Log n: - £ Log NjJ)
H
J — ^max
1=1
l/N Log N!
J' =
21 A
log S — - 2^f
i-1
S pi Log pA
S
-£ l/s Log l/S
lnl
-68-
-------
During the first year of sampling, 10 to 15 two-minute trawl tows were taken on
each of the inner stations. Species per unit area, ENS, H, and H' for each tow were
plotted (as a % of the totals for 10 trawl tows) against sample number for each sta-
tion. All 4 stations showed comparable results. This approach was used to determine
the degree of efficiency of the highly variable trawling technique (Taylor, 1953) so
that the least number of replicates would be taken for comparative purposes. Most
of the curves became asymptotic by the 7th sample; on this basis, 7 tows were taken
for the remainder of the sampling period. A more complete analysis of this method is
in preparation. Additional work has shown that repetitive sampling is more efficient
than single trawl methods. A comparison of seven 2-minute tows with one 14-minute tow
taken simultaneously once every 4 hours over a 24 hour period showed that the repetitive
technique accounted for 59.4% more individuals and 31.8% more species of fishes than the
single trawl method. Although the otter trawl is limited with respect to the numbers
and types of organisms sampled, by calibrating this method in terms of trawlable organ-
isms taken per unit effort, a valid comparison between contaminated and control stations
was feasible.
The CXindex of overlap (Morisita, 1959; Horn, 1966) was used to determine the ex-
tent of faunal similarity among the offshore sites. The CA index represents the proba-
bility that 2 randomly drawn individuals from populations x and y (see below) will be
of the same species relative to the probability that 2 individuals of the same species
will be drawn from population x or y alone (Home, 1966).
S
cA =
i=l
X X =
Xy =
.2
2
(Ax +>y) XY
x
y
-69-
-------
where S = Number of species
xi and yi = numbers of individuals of the
ith species in populations X and Y
X and Y = total number of individuals in tb-
2 samples
Xx and Xy = measures of diversity
(Simpson, 1949) as modified for sampling
with replacement (Horn, 1966).
Annual total numbers of fishes collected at each station were used to minimize monthly
variability; the cA indices were calculated for all the combinations of inner, outer,
and transect stations. Transect data were analyzed separately from the monthly sta-
tions. The cAindex equals 0 when the samples have no species in common and approxi-
mately 1 when the samples are the same with respect to the proportionality of species
composition.
In addition to the use of monthly and annual total numbers of individuals (N) and
(S-1)
numbers of species (S), a species richness component (SR) was calculated as SR = ^0g ^
(Margalef, 1958; Boesch, 1973). Dominance-diversity curves were constructed on annual
totals for each station by plotting the Log of the importance value (N) against the
importance rank of each species. Dominance values were patterned after Berger and
Parker (1970) and McNaughton (1968) as follows:
A -R
»1 +«2
"»= „
N1 + k2 4 n3
Dc =
N
where N , N_, and N_ = the numbers of individuals of the first, second, and third most
1 * V
abundant species (respectively), and N = the total number of individuals.
-70-
-------
Table 1 . Comparison of Different Parameters Relative to Significant Levels of Variation (p<.01) Between Comparable
Stations ( ~ = Significant dif£erence;0 = No Significant Variation). Mean Values Per Month Over a Period
of One Year Were Used to Determine significance.
Parameter
Pair
1
(E7, F9)
2
(E8.F10)
3
(E9,F16)
4
( E10,F11)
5
(E12.F12)
6
(Ell,F14)
7
(E13.F15)
Number of
Species (S)
+
+
0
+
0
+
+
Number of
Individuals (N)
+
+
0
~
0
+
+
Species
Richness (SR)
+
0
0
+
0
0
0
Expected Number
of Species
(ENS)
~
0
0
+
0
0
0
-------
A computerized statistical analysis was made to determine the relationships
of the corresponding (paired) stations of the study areas. Monthly data from
June, 1972 to May 1973 were used; this included number of species (S), number
of individuals (N), species richness (SR), and expected number of species (ENS).
An analysis of variance (repeated measures design) was used on the paired
stations, and a least significant difference test allowed a comparison of the
annual averages of the paired stations (Table 1).
RESULTS
Physical and Chemical Data
Mean temperatures of the grouped offshore stations in the 2 sampling areas
are shown in Fig. 2. Peaks usually occurred in August and September while the
lowest temperatures were found during January and February. Periods of rapid
change were evident during fall (September-November) and spring (March-May)
months. Variable winter temperatures during the second year of sampling re-
flected air temperature changes due to cold fronts moving through the area.
Water temperatures in shallow coastal areas of north Florida follow the air
temperature rather closely (Koenig «it al., 1974). Water temperatures were
comparable in the 2 study areas. Color data are shown in Fig. 3. Fenholloway
River stations below the pulp mill (F3-F9) showed substantially higher color
levels than Fenholloway stations above the pulp mill or comparable Econfina
sites. The inshore Fenholloway stations (Fg, F^q, F^, Fj^) were consistently
higher in color than their Econfina counterparts. There were periodic in-
creases in color in both systems during winter and early spring months; salinity
and color had a generally inverse relationship (Fig. 3), thus implicating land
runoff with color increases in offshore areas. Particularly heavy rainfall
during the spring of 1973 produced extremely low salinity and high color
conditions on all stations. The Fenholloway River stations had high turbidity
-72-
-------
and low sccchi disk readings below the pulp mill when compared with stations
above the mill or with their Econfina River counterparts (Fig. 4). The
inshore Fenholloway stations (Fg, F^q, F^, F^) were extremely turbid; any
seasonal increases during winter and early spring were possibly masked by the
generally high levels. The entire offshore Fenholloway area had consistently
higher turbidity and lower light penetration than the Econfina system. Student-t
tests run on mean values of color and turbidity between comparable stations
of the two systems showed significant differences Ha^F^-^E'
P ^ 0.05) between each pair except Eg and (Lewis, 1974). The general
increase in the color and turbidity of the Fenholloway system was thus considered
to be a direct result of the pulp mill operation. This effect was found in
extensive offshore areas with recovery occurring in the vicinity of station Fi6»
Representative dissolved oxygen data are shown in Figs. 5 and 6. Below
the pulp mill, there was little or no oxygen in the Fenholloway River at var-
ious times of the year, although there was a relatively rapid recovery in the
Fenholloway estuary (stations Fj, Fg, Fg). This was attributed to mixing of
the anoxic river with water from the Gulf of Mexico. The dissolved oxygen con-
tent of the Econfina River was usually below that of the open Gulf; this could
be attributed to the introduction of spring-fed water of low oxygen content into
this river system although the exact cause of this phenomonen remains unknown.
During the summer, there were considerable differences in the (benthic) dissolved
oxygen content of the Econfina and Fenholloway offshore stations. Readings that
were substantially less than 50% saturation were found on all stations with
the exception of station F^* Such reduced D.O. levels could be attributable
to higher microbial activity in the Fenholloway sediments as well as differences
in the level of oxygen production by benthic plants (Livingston et al., 1972;
Zimmerman, 1974). The diurnal fluctuation of dissolved oxygen in both areas
(high during the day, low at night) was not unusual considering the time of
year and location of sampling sites.
-73-
-------
Biological Data
Fresh water fish collections (rotenone, seines, dip nets) made before and
after the operation of the pulp mill showed that the Fenholloway River does
not have a normal fish fauna in areas affected by PME (Livingston et al., 1972).
Only 3 species were found during the study compared with nearly 30 species
collected in control areas. Most of the fishes taken in the Fenholloway were
topmlnnows that could have been washed into the river from surrounding areas.
In any case, the river was obviously subject to acute effects of PME, and was
thus not intensively monitored during this study except for those parameters
that were considered meaningful to the impact on the estuarine and coastal
systems. Except for occasional movements of a few species in the salt wedge,
there was no permanent fish assemblage found in the Fenholloway estuary.
Collections from the brackish water marshes of the 2 river systems are shown
in Table 2. After 18 months of collections, 2,789 individuals of 48 species
were taken from the Econfina marsh compared to 34 individuals of 6 species from
the Fenholloway marsh. It would appear that the Fenholloway marshes were
affected by the pulp wastes. Except for periodic influxes of small numbers of
topmlnnows from surrounding areas, no permanent fish fauna was established.
Although these areas were occupied by a seemingly undisturbed growth of needle-
rush (Juncus roemerianus) and other marsh vegetation, none of the usual inverte-
brates such as fiddler crabs (Uca spp) were found*
During the period of collection, several extensive fish kills were observed at
the mouth of the river. These data would thus indicate that due to the influx
of PME and associated environmental conditions, the Fenholloway River-Estuary
system was depauperate with respect to numbers of individuals and species of
fishes.
-74-
-------
Table 2 : Fishes collected monthly in the marsh
1 8 month
areas of the
period (1971-
Fenholloway
¦1972)
and Econfina systems during
an
Fenholloway System:
Marsh Station - (5b)
1971
April
10
1971
Hay
6
1971
June
12
1971
July
13,27,29
1971
Aug.
23
1971
Sept.
16
1971
Oct.
14
1971
Nov.
19
1971
Dec.
22
Poecilia latiplnna
N
N
7
N
N
N
1
N
N
Gambusia affinls
0
0
4
0
0
0
0
0
Fundul us plmlUs
F
F
3
F
F
F
F
F
Fundulus confluentus
I
I
I
I
I
1
I
I
Fundulus grandis
S
S
S
S
S
1
S
S
Jordanella floridae
H
H
H
H
H
1
H
H
1972
Jan.
22
1972
Feb.
12
1972
March
11
1972
April
20
1972
May
18
1972
June
8
1972
July
3
1972
Aug.
22
1972
Sept.
9
Poecilia latiplnna
N
N
N
3
6
N
0
N
0
N
0
N
0
Gambusia affinls
0
0
0
3
3
Fundulus confluentus
F
F
F
1
F
F
F
F
Fundulus grandis
I
I
I
I
I
I
I
Jordanella floridae
S
S
S
S
S
S
S
H
H
H
H
H
H
H
-------
Table 2: Marsh Station - (4a)
1971
April
17
1971
May
20
1971
June
22
1971
July
23
1971
August
10
1971 1971
September October
29 14
1971 1971
November December
7 9
Leiostomus xanthurus
103
50
42
43
2
4 5
8
Lucanla parva
1
1 2
Anftuilla rostrata
3
Ictalurus catus
2
La godon rhomboides
13
3
10
10
16
Strorgylura marina
1
1
2
3
1
Menidia beryllina
30
9
68
126
56
22 10
4 2
Trinectea maculatus
1
2
2
Gobiosotna bosci
13
2
5
6 1
1
Elops saurus
2
Myrophis punctatus
9
14
9
8
4
Musil cephalus
6
4
5
Harengula pensacolae
1
Fundulus similis
2
9
9 6
25 4
Brevoortia patronus
66
62
18
270
Anchoa mitchilli
61
50
80
1
Lepisosteus osseus
1
-------
Table 2' Marsh Station - (4a) continued
1971
1971
1971
1971
1971
1971
1971
1971
1971
April
May
June
July
August
September October
November
December
17
20
22
23
10
29
14
7
9
Eucinostomus argenteus
1
13
23
10
9
Cynoscion nebulosus
2
4
Poecilia latipinna
6
105
52
11
Olieoplites saurus
1
Arius fells
2
Microgobius gulosus
1
2
Archosaraus probatocephalus
4
Paralichthys lethostigraa
1
Fundulus erandis
110
30
95
13
Fundalus confluentus
3
15
Adinia xenica
15
16
10
Gambusia affinis
3
7
Jordanella floridae
5
Svmphurus plagiusa
1
Prionotus tribulus
1
1
Cyprinodon variegatus
12
2
8
9
Mugil curema
4 12
-------
Table 2: Marsh Station - (4a) continued
1971 1971 1971 1971 1971 1971 1971 1971 1971
April May June July August September October November December
LZ 20 22 23 10 29 14 7 9
Monacanthus hispidus 2
Ancyclopsetta quadrocellata
Bairdiella chrysura
Ophichthus Romes^
Eucinostomus gula
Labidesthes sicculus
Caranx latus
Dasyatis sabina
co
i
Cynoscion arenarius
Strongylura timucu
Sardinella anchovia
Anchoa hepaetus
Dorosoma robustym
-------
Table 2 : Econfina Marsh Station - (4a) continued
1972
1972
1972
1972
1972
1972
1972
1972
1972
January
February
March
April
May
June
July
August
September
29
26
30
27
27
12
11
24
14
Leiostomus xanthurus
13
8
4
Lucania parva
12
12
Anguilla rostrata
Ictalurus catus
Lagodon rhomboidea
Strongylura marina
Menidia beryllina
18
25
11
VO
I
Trinectes maculatus
Gobiosoma boaci
Elops aaurus
Mvrophis punctatua
Mugil cephalus
Harengula pensacolae
Fundulus aimilis
Brevoortia patronus
Anchoa mitchilli
16
114
142
Lepisosteus osseus
Eucinostomus argenteus
5
-------
Table 2: Econfina Marsh Station - (4a) continued
1972
January
29
1972
February
26
1972
March
30
1972
April
27
1972
May
27
1972
June
12
1972
July
11
1972
August
24
1972
September
14
Cynoscion nebulosus
6
Poecilia latipinna
10
24
Oligoplites saurus
11
Arius felis
1
Microgobius gulosus
Archosargus probatocephalu9
Paralichthvs lethostigma
Fundulus grandis
14
1
3
Fundulus confluentus
3
Adenia xenica
16
1
1
3
12
Gambusia affinis
162
Jordanella floridae
Symphurus plagiusa
Prionotus tribulus
Cyprinodon variegatus
10
Mugil curema
13
Monacanthus hispidus
-------
Table 2: Ecoufina Marsh Station - (4a) continued
1972
1972
1972
1972
1972
1972
1972
1972
1972
January
February
March
April
May
June
July
August
September
29
26
30
27
27
12
11
24
14
AncvcloDsetta auadrocellata
2
Bairdiella chrvsura
1
Oohichthus gomesi
1
Eucinostomus eula
Labidesthes sicculus
1
9
31
Caranx latus
5
Dasyatis sabina
I
2
Cynoscion arenarius 1
Strongylura timucu 1
Sardinella anchovia
11
Anchoa hepsetus
Dorosoma robus turn
-------
A comparison of the numbers of individuals (N) and species (S) taken
during repetitive sampling of the inner stations (Ey, Eg, F^, F^q) is shown in
Fig. 7. Fall and spring peaks (N) were evident at all 4 stations. Fenholloway
stations had consistently lower numbers of fishes than their Econfina counter-
parts although such differences were minimal during the first winter (January-
March, 1972) of sampling and nonexistent during the second winter (December-
February, 1973). The numbers of fishes taken from the Fenholloway stations
wer/e thus substantially lower than the control area although such differences
were seasonally controlled. The Fenholloway S values were generally lower than
their Econfina counterparts, with the spring-fall pattern not as uniform as
that of the N curves. Summed data for the 5 outer stations of each system
showed similar reductions in numbers of individuals. A detailed (monthly)
breakdown of the N and S data for all stations during the final year of trawl-
ing is shown in Fig. 8. Peaks (N and S) occurred during the fall and spring
with seasonal changes in the numbers of individuals and species roughly coin-
ciding with periodic temperature changes. Except for Fj^ an^ N values of
the Fenholloway stations were substantially less than their Econfina counter-
parts. However, with the possible exception of the annual (total) S values
for both systems were comparable. There was little variation in the monthly
values of S among the Econfina stations; with the exception of F^ and
(which resembled the Econfina stations), there was a uniform reduction of the
monthly Fenholloway S values when compared to the control stations. The average
numbers of individuals and species taken per month on the Fenholloway system as
a whole was significantly less (P <.01) than those taken in the control area.
The reductions of N and S at stations Fg, F^q, ®"ll» Fi4> an^ F^ were statisti-
cally significant (Table 1). These data would thus indicate that although the
(cumulative) total number of species caught at each Fenholloway station did
-82-
-------
not differ substantially from its Econfina counterpart, the monthly values
of S at the above 5 stations were significantly less than the control values.
In terms of N and S, such reductions were compatible with the physico-chemical
data.
Table 3 shows the cumulative species composition of the offshore stations.
A small group of species dominate in both areas with the pinfish (Lagodon
rhomboides) as the most prevalent species. This was evident in the offshore
Econfina stations (Eg, Eg, E^g, E13) and the least affected Fenholloway
stations (Fi2> F^g). T^ie inshore Econfina stations (Ey, E^q, E^) had consider-
able numbers of spot (Leiostomus xanthurus) and/or silver perch (Bairdiella
chrysura) while the stations farther offshore included such dominants as
fringed filefish (Monacanthus ciliatus), southern sea bass (Centropristis
melarta), and spottail pinfish (Diplodus holbrooki). The Fenholloway inshore
stations (Fg, F^, F^) were dominated by the bay anchovy (Anchoa mitchilli).
Other Fenholloway stations (F^Qj *"15) had considerable numbers of spot and
pinfish. In general, the species distribution reflected proximity to the river
mouth, distance from shore, and the general level of PME. From a total number
of 73 species, 45 were common to both systems. Each drainage area had 14
species that were not found in the other. Dominance diversity curves (Fig. 9)
and dominance values (Table 4) amplify these results. The most pronounced
dominance was found on the inshore Econfina stations (Ey, E^q, Ej^) and
The offshore areas (Eg, F^) had relatively low dominance values as did most of
the Fenholloway region (F9, F10, F12, F14, F^). Thus, with the exception of
^11' where the bay anchovy was predominant, areas that were found to be under
stress from PME had consistently lower dominance than the corresponding control
areas.
At most stations, there was a (seasonal) pattern of species diversity (ENS)
and species richness (SR) (Fig. 10). Peaks usually occurred in the fall
-83-
-------
Table 3:
ECONFINA SYSTEM (June, 1972 - May, 1973)
Station 7
Leiostonus xanthurus 685
Lagodon rhomboides 110
Bairdiella chrysura 84
Polydactylus octonemus 70
Eucinostomus gula 44
Monacanthus hispidus 33
Chasmodes saburrae 28
Opsanus beta 24
Syngnathus scovelli 16
Gobiosoma robustum 14
OrthopristLs chrysoptera 11
Urophycis floridanus 11
Microgobius gulosus 11
Monacanthus ciliatus 15
Sphoeroides nephelus 9
Centropristis melana 8
Calamus arctifrons 8
Syngnathus floridae 7
Cynoscion nebulosus 6
Chilomycterus schoepfl 5
Chloroscorabrus chrysurus 5
Hippocampus erectus 4
Diplodus holbrooki 3
Archosargus probatocephalus 3
Chaetodipterus faber 3
Hypleurochilus geminatus 3
Micropogon undulatus 3
Trinectes maculatus 3
¦Symphurus plaguisa 3
Caranx hippos 2
Anchoa mitchilli 2
Arius felis 2
Lutjanus griseus 2
Prionotus tribulus 2
Cyprinodon variegatus
Hippocampus zosterae
Hypsoblennius hentzi
Lactophrys quadricornis
Lepisosteus osseus
Paraclinus fasciatus
Synodus foetens
Paralichthys albigutta
Bathygobius soporator
Station 8
Lagodon rhotnboides 629
Monacanthus ciliatus 165
Bairdiella chrysura 124
Syngnathus floridae 98
Monacanthus hispidus 87
Diplodus holbrooki 71
Centropristis melana 66
Orthopristis chrysoptera 49
Leiostomus xanthurus 41
Opsanus beta 41
Eucinostomus gula 33
Chilomycterus schoepfi 24
Urophycis floridanus 16
Syngnathus scovelli 14
Calamus arctifrons 13
Cynoscion nebulosus 12
Paraclinus fasciatus 11
Sphoeroides nephelus 8
Haemulon aurolineatum 7
Anchoa mitchilli 6
Polydactylus octonemus 5
Symphurus plagiusa 5
Syngnathus louisianae 5
Chasmodes saburrae . 4
Synodus foetens 4
Aluterus schoepfi 3
Gobiosoma robustum 3
Micrognathus crinigerus 3
Paralichthys albigutta 3
Chaetodipterus faber 2
Eucinostomus argenteus i
Hippocampus erectus 1
Paraclinus marmoratus 1
Paralichthys lethostigma
Diplectrum formosum
Prionotus scitulus
Lactophrys quadricornis
Mycteroperca microlepis
Trinectes maculatus
Station 9
lagodon rhomboides 300
Monacanthus ciliatus 172
Centropristis melana 146
Calamus arctifrons 141
Orthopristis chrysoptera 133
Monacanthus hispidus 98
Bairdiella chrysura 55
Syngnathus floridae 49
Diplodus holbrooki 38
Chilomycterus schoepfi 24
Opsanus beta 20
Haemulon aurolineatum 19
Sphoeroides nephelus 15
Lactophrys quadricornis 14
Diplectrum formosum 9
Synodus foetens 9
Eucinostomus gula 8
Urophycis floridanus 7
Cynoscion nebulosus 6
Aluterus schoepfi 3
Gobiosoma robustum 3
Micrognathus crinigerus 3
Paraclinus fasciatus 3
Paralichthys albigutta 3
Syngnathus louisianae 3
Chaetodipterus faber 2
Hippocampus erectus 2
Leiostomus xanthurus 2
Ancyclopsetta quadrocellata
Sygnathus scovelli
Hippocampus zosterae
Lutjanus griseus
Mycteroperca microlepia.
Nicholsina usta
Ogcocephalus radiatus
Peprilus burti
Scorpaena grandicornis
-84-
Station 10
Lagodon rhomboides 1006
Bairdiella chrysura 155
Monacanthus ciliatus 124
Syngnathus floridae 79
Diplodus holbrooki 72
Eucinostomus gula 64
Orthopristis chrysoptera 63
Centropristis melana 51
Leiostomus xanthurus 48
Monacanthus hispidus 42
Eucinostomus argentus 3 2
Opsanus beta 25
Cynoscion nebulosus 24
Gobiosoma robustum 18
Calamus arctifrons 13
Chilomycterus schoepfi 13
Syngnathus scovelli 12
Micrognathus crinigerus 9
Haemulon aurolineatum £
Urophycis floridanus 8
Paraclinus fasciatus 7
Harengula pensacolae 5
Sphoeroides nephelus 4
Syngnathus louisianae 4
Diplectrum formosum 3
Hippocampus erectus 3
Lactophrys quadricornis 3
Paralichthys albigutta 3
Aluterus schoepfi
Lutjanus griseus
Synodus foetens
Trinectes maculatus
Chaetodipterus faber
Chasmodes saburrae
Cholorscombrus chrysurus
Dasyatis sabina
Hippocampus zosterae
Microgobius gulosus
Fundulus grandis
-------
TABLE 3:
ECONFINA SYSTEM (June 972 - May, 1973)
Station 11
§
Station 12
3
Station 13
#
Lagodon rhomboides
621
Monacanthus ciliatus
416
Lagodon rhomboides
471
Leiostomus xanthurus
205
Lagodon rhomboides
368
Centropristis melana
168
Bairdiella chrysura
125
Dipoldus holbrooki
163
Bairdiella chrysura
138
Eucinostomus gula
125
Bairdiella chrysura
138
Diplodus holbrooki
96
Monacanthus ciliatus
84
Syngnathus floridae
135
Monacanthus hispidus
92
Monacanthus hispidus
58
Centropristis melana
76
Monacanthus ciliatus
83
Centropristis melana
45
Calamus arctifrons
66
Leiostomus xanthurus
78
Orthopristis chrysoptera
43
Monacanthus hispidus
65
Orthopristis chrysoptera
76
Syngnathus floridae
43
Orthopristis chrysoptera
51
Eucinostomus gula
65
Opsanus beta
42
Chilomycterus schoepfi
36
Calamus arctifrons
57
Urophycis floridanus
24
Eucinostomus gula
34
Syngnathus floridae
47
Cynoscion nebulosus
19
Urophycis floridanus
13
Gobiosoma robustum
41
Diplodus holbrooki
15
Cynoscion nebulosus
9
Paraclinus fasciatus
41
Polydactylus octoneraus
14
Syngnathus louisianae
9
Opsanus beta
23
Gobiosoma robustum
12
Haemulon aurolineatum
8
Chilomycterus schoepfi
21
Chasmodes saburrae
10
Opsanus beta
8
Urophycis floridanus
12
Chaetodipterus faber
9
Lactophrys quadricornls
7
Haemulon aurolineatum
9
Chilomycterus schoepfl
9
Leiostomus xanthurus
6
Chasmodes saburrae
6
Microgobius gulosus
9
Sphoeroides nephelus
5
Paralichthys albigutta
6
Paraclinus fasciatus
8
Micrognathus crinigerus
5
Aluterus schoepfi
5
Syngnathus scovelli
8
Syngnathus scovelli
5
Hippocampus erectus
5
Sphoeroides nephelus
7
Aluterus schoepfi
4
Anchoa mitchilli
4
Paralichthys albigutta
5
Gobiosoma robustum
4
Polydactylus octonemus
4
Calamus arctifrons
4
Paraclinus fasciatus
3
Micrognathus crinigerus
4
Microglia thus crinigerus
4
Paraclinus marmoratus
3
Cynoscion nebulosus
3
Syngnathus louisianae
4
Diplectrum formosum
2
Diplectrum formosum
2
Synodus foetens
3
Hippocampus erectus
2
Sphoeroides nephelus
2
Anchoa mitchilli
2
Hippocampus zosterae
2
Lactophrys quadricornls
2
Cyprinodon variegatus
2
Paralichthys albigutta
2
Syngnathus louisianae
2
Lactophrys quadricornls
2
Synodus foetens
2
Chaetodipterus faber
Arius felis
1
Gymnothorax nigromarginatus
1
Hippocampus zosterae
Gymnothorax nigromarginatus
1
Harengula pensacolae
1
Micropogon undulatus
Hippocampus zosterae
1
Mugil cephalus
1
Syngnathus scovelli
Hypleurochilus geminatus
1
Ogcocephalus radiatus
1
Synodus foetens
Hypsoblennlus hentzi
1
Trinectes maculatus
Lutjanus griseus
1
Apogon townsendl
Trinectes maculatus
1
Harengula pensacolae
-85-
-------
Tab le 3 '•
FENHOLLOWAY SYSTEM (June, 1972 - May, 1973)
Station 9
Anchoa mltchilli 114
Leiostomus xanthurus 59
Polydactylus octonemu9 48
Lagodon rhomboides 40
Eucinostomus gula 34
Cynoscion arenarius 24
Cynoscion nebulosus 22
Bairdiella chrysura 21
Centropristis melana 15
Orthopristis chrysoptera 7
Urophycis floridanus 7
Arius felis 6
Calamus arctifrons 6
Chaetodipterus faber 5
Monacanthus ciliatus 4
Chilomycterus schoepfi 3
Micropogon undulatus 3
Syngnathus floridae 3
Chloroscombrus chrysurus 2
Diplodus holbrooki 2
Menticirrhus saxatilis 2
Monacanthus hispidus 2
Dasyatis sabina 2
Sphoeroides nephelus 2
Ogcocephalus radiatus 2
Syngnathus scovelli
Anchoa hepsetus
Echeneis naucrates
Etropus crossotus
Lutjanus griseus
Opsanus beta
Faraclinus fasciatus
Paralichthys albigutta
Prionotus tribulus
Synodus foetens
Station 10
Leiostomus xanthurus 114
Lagodon rhomboides 60
Eucinostomus gula 45
Anchoa mitchilli 31
Urophycis floridanus 23
Monacanthus hispidus 18
Chilomycterus schoepfi 12
Bairdiella chrysura 11
Syngnathus floridae 10
Centropristis melana 9
Monacanthus ciliatus 9
Calamus arctifrons 7
Orthopristis chrysoptera 7
Arius felis 6
Synodus foetens 6
Sphoeroides nephelus 6
Cynoscion nebulosus 5
Caranx hippos 4
Diplectrum formosum 4
Gobiosoma robustum 4
Etropus crossotus 3
Syngnathus scovelli 3
Ogcocephalus radiatus 2
Opsanus beta 2
Paralichthys albigutta 2
Polydactylus octonemus 2
Chasmodes saburrae
Elops saurus
HaemuIon aurolineatum
Lactophrys quadricornis
Ogcocephalus radiatus
Prionotus tribulus
Selene vomer
Syngnathus louisianae
Dasyatis sabina
Station 11
Anchoa mitchilli 206
Leiostomus xanthurus 36
Polydactylus octonemus 34
Bairdiella chrysura 30
Eucinostomus gula 16
Chilomycterus schoepfi 15
Prionotus scitulus 9
Lagodon rhomboides 7
Arius felis 6
Caranx hippos 6
Prionotus tribulus 5
Etropus crossotus 4
Cynoscion nebulosus 3
Paralichthys albigutta 3
Centropristis melana 2
Chaetodipterus faber 2
Monacanthus ciliatus 2
Ogcocephalus radiatus 2
Synodus foetens 2
Chasmodes saburrae 2
Diplectrum formosum 1
Echeneis naucrates 1
Micropogon undulatus 1
Opsanus beta 1
Paralichthys lethostigma
Syngnathus floridae
Syngnathus scovelli
Urophycis floridanus
Station 12
Lagodon rhomboides 33 2
Centropristis melana 188
Bairdiella chrysura 169
Monacanthus ciliatus 143
Monacanthus hispidus 109
Orthopristis chrysoptera 104
Leiostomus xanthurus 100
Calamus arctifrons 46
Diplodus holbrooki 46
Syngnathus floridae 33
Haemulon aurolineatum 30
Chilomycterus schoepfi 23
Urophycis floridanus 21
Anchoa mitchilli 19
Eucinostomus gula 12
Cynoscion nebulosus 11
Sphoeroides nephelus 9
Paraclinus fasciatus 7
Diplectrum formosum 4
Opsanus beta 4
Polydactylus octonemus 4
Paralichthys albigutta 3
Prionotus scitulus 3
Syngnathus louisianae 3
Harengula pensacolae 3
Synodus foetens 2
Menidia beryllina 2
Chasmodes saburrae 1
Dasyatis sabina 1
Etropus crossotus 1
Gymnothorax nigromarginatus
Lutjanus griseus
Nicholsina usta
Sphyraena borealis
Hippocampus erectus
Microgobius gulosus
Table 3: Qualitative list of fishes taken in repetitive trawl-tows (11 month totals for "outer" stations; 12 month
totals for "inner" stations) of individuals ranked by numbers
-------
Table
FENHOLLOWAY SYSTEM (Jv , 1972 -
Station 14
Anchoa mitchilli
Polydactylus octonemus
Leios Coraus xanthurus
Lagodon rhomboides
Bairdiella chrysura
Orthopristis chrysoptera
Caranx hippos
Eucinostomus gula
Urophycis floridanus
Gobiosoma robustum
Monacanthus hispidus
Chasmodes saburrae
Chaetodipterus faber
Cynoscion nebulosus
Etropus crossotus
Sphoeroides nephelus
Syngnathus scovelli
Centropristis melana
Synodus foetens
Menticirrhus americanus
Arius felis
Calamus arctifrons
Chilomycterus schoepfi
Chloroscombrus chrysurus
Ogcocephalus radiatus
Paralichthys albigutta
Dasyatis sabina
Diplectrum formosum
Diplodus holbrooki
Lutjanus griseus
Monacanthus ciliatus
Opsanus beta
Prionotus tribulus
Syngnathus floridae
Microgobius gulosus
Harengula pensacolae
Hippocampus erectus
Station 15
Leiostomus xanthurus
Lagodon rhomboides
Eucinostomus gula
Monacanthus ciliatus
Chilomycterus schoepfi
Syngnathus floridae
Orthopristis chrysoptera
Centropristis melana
Synodus foetens
Urophycis floridanus
Monacanthus hispidus
Anchoa mitchilli
Paralichthys albigutta
Diplectrum formosum
Mycteroperca microlepis
Polydactylus octonemus
Syngnathus louisianae
Opsanus beta
Sphoeroides nephelus
Aluterus schoepfi
Microgobius gulosus
Calamus arctifrons
Caranx hippos
Lactophrys quadricornis
Arius felis
Syngnathus scovelli
Bairdiella chrysura
Ancyclopsetta quadrocellata
Astroscopus y-graecum
Cynoscion nebulosus
Etropus crossotus
Haemulon aurolineatum
Hippocampus erectus
Micrognathus crinigerus
Paralichthys lethostigma
Prionotus tribulus
Prionotus scitulus
Astrapogon stellatus
Diplodus holbrooki
#
115
57
46
30
28
12
11
11
8
7
6
6
4
4
4
4
4
3
3
2
2
2
2
2
2
-87-
May, 1973)
#
Station 16
#
65
Lagodon rhomboides
164
58
Monacanthus ciliatus
158
45
Bairdiella chrysura
133
24
Centropristis melana
78
16
Calamus arctifrons
77
14
Orthopristis chrysoptera
76
13
Monacanthus hispidus
65
11
Syngnathus floridae
37
11
Chilomycterus schoepfi
27
11
Eucinostomus gula
22
8
Urophycis floridanus
19
6
Diplodus holbrooki
16
6
Cynoscion nebulosus
14
5
Leiostonus xanthurus
11
5
Lactophrys quadricornis
12
5
Haemulon aurolineatum
8
5
Anchoa mitchilli
6
4
Diplectrum formosum
6
4
Prionotus scitulus
6
3
Nicholsina usta
5
3
Synodus foetens
4
3
Opsanus beta
4
3
Aluterus schoepfi
3
3
Sphoeroides nephelus
3
2
Hippocampus erectus
5
2
Hippocampus zosterae
2
2
Paraclinus fasciatus
2
1
Paralichthys albigutta
2
1
Polydactylus octonemus
2
1
Arius felis
2
1
Gymnothorax nigromarginatus
1
1
Ogcocephalus radiatus
1
1
Prionotus tribulus
1
1
Scorpaena grandicornis
1
-------
Table 4: Dominance values for fishes taken on all offshore stations.
E7 Eg Eg E1q E]_2 En E13
Da
54.8
40.3
23.1
52.6
25.2
39.6
30.0
Db
63.7
50.9
36.4
60.7
47.4
52.7
40.7
70.4
58.8
47.6
67.2
57.3
60.7
49.5
F9
F10
f16
F11
F12
F14
h"1
cn
1
1
1
Da
25.4
27.3
20.2
51.1
23.1
29.6
18.5
%
38.9
41.7
35.7
60.4
36.0
44.3
34.9
Dc
47.5
52.5
49.0
68.8
47.8
52.1
42.7
-88-
-------
(September-October) and/or the spring (March-April). This pattern was not
evident at stations Eg and Fjg. Consistent reductions of both parameters
relative to the control values were noted at F and there were periodic re-
ductions at Fg and Fio. Mean values (Table 5) clearly show these relationships,
and statistical analysis (Table 1) has confirmed the significance of these
observations. Additional comparisons of other types of diversity indices
(Table 6) tend to support these data. Species diversity and species
richness indices were thus useful as indicators of the impact of PME at only the
most heavily stressed stations (Fg, F^). At others, such as F^q, F14 and F15,
where there were significant reductions of N and S, there was little or no
effect on the diversity parameters. Qualitative differences in species com-
position as well as dominance values appear to be related to these observations.
Transect data are shown in Figs. 11 and 12. The Fenholloway area (F^,
Fio» ^15) was characterized by substantial reductions in N and S values. How-
ever, peaks of both parameters occurred periodically on either side of this
area. In terms of numbers of individuals, this was especially pronounced at
station T21. Seasonal variations of this pattern are possible although the
limited sampling regime disallowed a rigorous test of such periodic changes.
The general pattern was consistent with respect to the reduced numbers of
species and individuals in the Fenholloway system. The peaks of N and S in
areas between the Fenholloway and Econfina drainages were more pronounced and
consistent than their counterparts in the Fenholloway Spring-Warrior Creek
transition area. The various diversity indices were well correlated
with relatively few differences between the scaled and unsealed data. The
substantially reduced salinities of the spring of 1973 could have caused the
low S values in the Econfina area; this was also reflected in the diversity
figures. There were distinct peaks of diversity on either side of the Fen-
holloway drainage during the July and November samplings. Such peaks were not
-89-
-------
Table 5: Mean values (1 year) of ENS S.R., and number of species taken per
month for paired outer stations
ENS SR S/month ENS SR S/month
e7
6.37
2.50
11.0
Fg
A.90
1.73
6.7
e8
6.65
2.58
12.8
F10
5.70
2.05
8.2
e9
7.85
2.69
12.8
*16
8.02
2.67
12.1
E10
7.46
2.37
12.0
Fll
4.62
1.56
5.3
Ei 2
6.72
2.37
12.5
F12
7.09
2.53
12.3
E11
5.87
2.36
11.9
f14
5.91
1.88
7.0
e13
7.32
2.63
13.3
F15
6.81
2.26
8.3
-90-
-------
Table 6: Comparison of BriHouin (H) and Shannon-Wever (H') diversity indices for 2 sets of stations
(E7, Eg, F9, Fiq)
e7
^8
v
*10**
Date
H
H1
Date
H
Hi
Date
H
11
Date
H
Ml
July,'71
1.18
1.26
June,'71
1.33
1.45
July,'71
1.77
1.95
June,'71
1.78
2.21
Dec.,'71
1.78
1.94
Dec.,'71
1.63
1.80
Sept.,71
1.63
1.74
Sept.,'71
1.11
1.36
Jan., 1 72
1.41
1.60
Jan.,'72
1.48
1.77
Nov.,'71
0.58
0.60
Nov.,'71
1.52
1.91
Feb. , ' 72
0.67
0.71
Feb.,'72
1.61
1.76
March,'72
1.28
1.51
April,'72
1.25
1.34
March,'72
1.52
1.70
March,'72
1.65
1.80
April,172
0.65
0.68
May,'72
1.19
1.61
April,'72
1.10
1.17
April,'72
1.30
1.35
May,'72
1.65
1.95
June,'72
1.55
2.00
May,'72
1.76
1.95
May,'72
1.63
1.86
June,'72
0.99
1.14
June,'72
1.85
2.02
June,'72
1.35
1.45
Mean
1.41
1.54
Mean
1.50
1.66
Mean
1.22
1.37
Mean
1.41
1.74
* Jan.,'72; Feb.,'72 = ^ 10 individuals
** Jan.,'72; Feb.,'72, March,'72, June,'72 = 10 individuals
-------
always correlated with the N and S increases. The evenness factor (J') was
relatively constant with little or no decrease in areas affected by PME. The
general patterns of transect N, S, and ENS distributions thus complement the
monthly data.
Trellis diagrams based on cX formulations for the transect and regular
stations are shown in Fig. 13 while a composite diagram of summarized data is
found in Fig. 14. There was a generally high level of similarity in the ichthyo-
faunal composition of various offshore stations. Natural (close) groupings were
evident on each side of the Fenholloway drainage (Eg, E^q, E^, E^ > T2p T2()»
and F10, F15, F^, T^, ^22^' T^ere were strong secondary relationships be-
tween these groups. Stations associated with the 3 rivers (E7, F9, T23) were
differentiated from the transect stations. The relatively high similarity of
the inshore Fenholloway stations (Fg, F]^, Fj^) could have reflected the impact
of PME on the area. The transect relationships reenforce the conclusion that
there are closely related fish assemblages on either side of the Fenholloway
drainage, and that these 2 groups are related (Ei3-T20_T22)' "^e offshore sta-
tions (E9, F^2» f16^ formed another natural group. These data were consistent
with the previously described qualitative and quantitative data on the fish
distribution in the 2 study areas, and allowed a quantitative reenforcement of
the conclusions based on qualitative distributions of the component fish species.
Overall, the various lines of data analysis converge to allow an assess-
ment of the effects of pulp mill effluents on the offshore fish fauna. In the
most severely affected areas, there were significant reductions of N, S, ENS,
H, H', and SR. Although species diversity indices were generally reduced in such
areas, there was little or no change in the equitability component. In areas of
transition between polluted and unpolluted systems, the relationships became far
more complex. Species diversity was no longer useful as a measure of Impact,
and actually showed moderate increases in transition zones. There were inter-
mediate areas that reflected substantial increases in the fish productivity. In
-92-
-------
general, there was a certain consistency in the quantitative and qualitative
aspects of the fish distribution that reflected not only the levels of the
kraft mill effluents, but also the various natural factors that are functions
of temporal as well as spatial mechanisms.
DISCUSSION
The data would indicate that pulp mill effluents have virtually destroyed
the fish fauna of considerable portions of the Fenholloway River and its
associated marshes. The Fenholloway estuary has also been severely affected
by the pulp wastes. An associated study comparing the bird populations in the
Fenholloway and Econfina marsh areas (Weiser, 1973) pointed out reduced numbers
of species and individuals of seasonal and permanent residents of the polluted
Fenholloway system. The reduced number of birds were attributed directly to the
reduced viability of the Fenholloway system, although some species were not
significantly affected. Shallow Gulf areas contiguous to the river mouth (within
2-3 km) were acutely affected by PME; there were reductions of numbers of fish
species and individuals, reduced species diversity, and a general decline of
water quality. Although there was a seasonal Influence on the degree of impact,
these data were somewhat comparable to the significant reductions of benthic in-
vertebrates due to the presence of bleached kraft mill effluents in a Canadian
estuary (Peer, 19/72); such reductions were attributed to long-term effects of
particulate material, flocculants, and/or toxic compounds. Associated dissolved
oxygen demand was noted, and this would correspond to summer D.O. reductions noted
in the Fenholloway area. ATP analysis (Quammen et al., 1971) indicated that pulp
mill wastes were confined to a small Inshore area within 1.7 km of the river mouth
and that such wastes were directed in a predominant SG direction due to oyster
reef placement and prevailing wind conditions. The fish distribution in the
area would suggest a somewhat broader area of acute impact. Inshore areas of
maximum effect occurred both west and southeast of the river mouth. In such
-93-
-------
shallow areas, both wind-induced and tidal currents were important in the
dissemination and impact of the pulp mill effluents.
Intermediate areas (between the unpolluted open Gulf and the effluent-
laden inshore waters) were exposed to chronic levels of PME; such levels
were largely determined by wind conditions and current patterns. These
transition areas represent a more complex biological condition than areas
of acute impact. Areas of chronic impact were marked by reduced numbers of
trawl-susceptible fishes, reduced numbers of species taken/month and lower
dominance values than the inshore stations. The repetitive collecting
technique allowed consistent results in the comparisons of paired stations.
Areas such as F 12 and F 16 remained relatively free of Impact. Of
particular interest, however, were the transect stations. The July series
indicates regions of reduced diversity at station 11 with increased diversity
in areas on either side of the Fenholloway drainage (T 21, T 19). A similar
diversity pattern prevailed in the November run, although there were
significant increases in numbers of individuals (T 21) and numbers of
species (T 21, T 19) in transitional areas. This pattern was repeated
during the April transect, although the diversity distribution at this
time was not clear; this condition was probably due, at least in part, to
decreased salinities in this area. McNulty (1961) described a narrow band
of increased numbers of benthic organisms a set distance from a sewage outfall
in Biscayne Bay (South Florida, U.S.A.). Such a "fertilizing" effect was
found to be a relatively stable (distance-related) parameter that was
characterized by certain "indicator" organisms in the areas of increased
productivity. Apart from certain obvious seasonal variations, one possible
explanation for this phenomenon in Apalachee Bay would be that the PME were
diluted in the transition areas to a point where enrichment due to increased
nutrients exceeded negative effects caused by toxicity, turbidity,
-94-
-------
sedimentation etc. Another possibility (not necessarily exclusive of the
first) would be that such areas reflected an edge effect between the polluted
(mud) communities of the Fenholloway system and the surrounding grass bed
communities. Such an area, as an ecotone, could shift due to seasonal or
irregular shifts in wind-induced movements of water or in response to
alt ei nLions of.' PMF1 levels Aliiumi'.h LiiCjf I no few d.Lla here l:o voijfy
such hypotheses, J conl i Diiat of LKi', -;i i:
-------
of Anchoa mitchllli, there was a reduced dominance at stations having
statistically significant reductions in N and S/month (F 9, F 10, F 14,
F 15). When compared to corresponding (control) stations, a general
pattern of reduced numbers of Individuals and species, decreased dominance,
and variable species richness and diversity was found in those Fenholloway
stations that were most affected by pulp effluent (as indicated by turbidity,
color, and secchi disk readings). However, except for F 11, the total number
of species taken (on an annual basis) at the Fenholloway stations did not
differ significantly from the control stations. Opportunistic species have
probably moved into stressed areas marked by lowered numbers of the usual
dominant species. Bechtel and Copeland (1970) considered that the bay
anchovy (Anchoa mitchilll) was the dominant species in areas stressed by
pollution although such increased pollution was not considered to have a
significant effect on the anchovy's food source. In this study, the bay
anchovy dominated at all 3 inshore stations that were under pollution stress
(F 9, F 11, F 14). However, only 2 of these stations had significant re-
ductions of species diversity. Since the bay anchovy is also a dominant
species in the relatively unpolluted (though highly turbid) Apalachicola
Bay (Livingston et al., 1974), it is possible - that this species is attracted
to areas of high turbidity regardless of the levels of pollution.
It is usually acknowledged that stressed conditions in estuarine
systems are related to low species diversity. Such areas are also character-
ized by high levels of dominance, and high productivity. It has been
pointed out (Boesch, 1972) that low species diversity is equated with
reduced structural stability (e.£.: low blocoenetic homeostatis). Odum
(1970) stated that because of low diversity, estuaries are thus particularly
susceptible to alteration by pollution with the reduction of susceptible
species and the empty niches claimed by less desirable species. Copeland
-96-
-------
(1970), on the other hand, claimed that low diversity systems that are
subjected to stress resist disturbances more easily than high diversity
systems adapted to constant environmental conditions. This latter hypothes
i8 supported by Boesch (1972) by indirect evidence (replacement of "typical
benthic species by eurytolerant ones in polluted areas). Dayton and
Hessler (1972), in an attempt to explain the relatively high species
diversity of deep-sea deposit feeders, developed the controversial
hypothesis that predictable cropping pressure reduces competitive exclusion
thus allowing high overlap of food resource utilization. Such a mechanism
would preclude high levels of dominance. On theoretical grounds, the role
of predation in this model has been established (Slobodkin, 1964; Connell,
1970; Paine, 1966; Dayton, 1971); with reduction of competitive exclusion
by cropping, there follows an enhanced opportunity for coexistence of
potential competitors and an increase of species diversity (increased
species richness and increased equitability). Although the estuarine
environment differs considerably from deep sea benthic conditions, the
principal considerations could be applied here. In the case of Apalachee
Bay, it is possible that the pollution acts as a non-specific predator
that "crops" (either by direct or indirect means) the various consti-
tuents in proportion to their abundance. If species density remains
low with food as a non-limiting resource, various (numerous) species
in low numbers can be found together with little competitive interac-
tion. As the chronic levels of pulp mill effluents reduce the dominant
species by one means or another, there could be a correlated increase
in cumulative numbers of species and species diversity. Regardless of
the exact mechanism involved, this study would indicate that species
diversity should not be used as the sole criterion of the impact of
-97-
-------
pollution on estuarine systems. The setting of guidelines for species
diversity levels that would indicate clean vs. polluted estuarine condi-
tions (Bechtel and Copcland, 1970; Holland £t al., 1973) would tend to
oversimplify the situation. It is possible that under certain circum-
stances, pollution can act as a predator. If this is true, the struc-
tural complexity of a system can remain intact even though the produc-
tivity is reduced and its population composition is altered. Swartz
ct^ a_l., (1973) found, also, that no single index is sufficient to indi-
cate water quality impact on estuarine and marine communities. Various
parameters such as faunal density, dominance, species richness, affinity
characteristics, and changes of population dynamics in spatial and tem-
poral terms are all necessary for the determination of the impact of a
pollutant on an estuarine community.
A number of associated studies have been carried out on the Econ-
fina-Fenholloway offshore system. Hooks (1973), working with trawlable
assemblages of invertebrates found significant decreases of invertebrates
in the Fenholloway area. Dominance (in terms of numbers of individuals)
was reduced in the Fenholloway system. While there were no significant
differences in the total numbers of species taken between the Fenholloway
and Econfina systems, the qualitative species composition did vary.
Species diversity (H, H1) and evenness (J, J1) comparisons were of limited
use in the analysis of this area. Heck (1973) verified these results
and postulated habitat alteration by kraft mill effluents as one of the
primary factors for the above observations. Due to a complex series of
changes involving siltation, turbidity, color, and associated toxic
effects, the benthic macrophyte assemblages in extensive portions of the
Fenholloway drainage system have been severely reduced (Livingston et al.,
1972; Zimmerman, 1974). In addition to qualitative differences in the
-98-
-------
species composition of the sea grasses and benthic algae over broad areas,
quantitative (dry weight) determinations showed a pronounced reduction of
t lie bcntliic mncrophyte assemblages compared to the (Econfina) control area.
These studies would indicate serious disruptions o£ the trophic relation-
ships in the Fenholloway area that could have direct application to the
observed fish distribution. It should be emphasized that the comparison
of tin: 2 systems is not completely satisfactory since it is possible that
natural differences between the two areas could contribute to some of
Lhesc observations. The presence of well formed oyster bars off the
mouth of the Fenholloway River would indicate that this is at least
partially true. However, the high level of correlation between PME
stress and impact at various levels of productivity would mitigate such
an objection. Because of the mobility of fishes and an acknowledged
(though uneven) tcndancy to avoid certain forms of pollution, it is
possible that the distribution of fishes in the Fenholloway system is
dependent in part on behavioral reactions to the PME. There is some
evidence that this could be true (Jones et al., 1956; Sprague, 1964;
Sprague, 1968; Sprague et^ al_., 1965; Ishio, 1965; Suiranerfelt and Lewis,
1967; Sprague and Drury, 1969; Hansen, 1969; Hansen and Matthews, 1972;).
Lewis (1974), working with KME from the pulp mill associated with the
Fenholloway system, found that Lagodon rhomboides and Fundulus grandis
lind threshold avoidance responses at 0.062% and 0.068% PME respectively.
Although no direct field application was available, such a response
would indicate that avoidance could be a contributing factor to the
observed distribution of these species in the Fenholloway system. It
is thus conceivable that the distribution of the fish fauna in the Fen-
«
holloway drainage area is dependent on multiple factors relating to the
-99-
-------
discharge of PME. In river, marsh, and adjoining estuarine areas, the
acute effects of such discharge have severely restricted the fish assem-
blages in terms of numbers of species and individuals. In the offshore
i.rcns, where chronic levels of PME prevail, various factors such as de-
creased benthic productivity, changes in numbers and distribution of
epibenlhic invertebrates, and avoidance reactions of fishes could con-
tribute to their observed distribution. These sublethal effects could
be related lo the species diversity and community structure of the af-
fected areas, although this connection remains obscure.
Although the seasonal distribution of larger fishes was not deter-
mined in this study, a survey was taken using gill nets, trammel nets,
and hook and line methods (night and day) to determine the top preda-
tors in the Econfina-Fenholloway portions of Apalachee Bay (Table 7).
In following the limited commercial and sports fisheries in the area,
it was found that the game and commercial species were reduced in num-
bers in an area roughly comparable to the area of chronic impact des-
/
cribed above for trawlable species. Although such data do not consti-
tute a quantitative verification of impact, it is of interest that the
larger fishes were distributed in a similar fashion to the trawlable
groups. The possible trophic relationships of such fishes and their
distribution with respect to the Fenholloway outfall, are presently
under study. In January, 1974, the pulp mill instituted a water pollu-
tion control program designed to eliminate considerable portions of the
biodegradable elements of the pulp mill effluent. The sampling program
will continue for 2 years beyond this date to determine the rate of re-
covery of the Fenholloway system. In addition to possibly solving some
of the questions related to trophic responses to the PME, it is hoped
that some of hypotheses developed during the initial stages of this study
will be tested as the bay system responds to the clean-up effort.
-100-
-------
TABLE 7.
Larftor Fishes Found in the Econfina-Fcnholloway Drainage Systems
Of Apalachee Bay at Night
Cnrchnrhinus leucas
CarchnrhJnus acronotus
Dasyatis sabina
Sphyrna tiburo
l.episoateus osseus
Elops saurus
Bagre marinus
Arlus felis
Paralichthys lethostigraa
Paralichthys albigutta
Rachycentron canadum
Cynoscion nebulosus
Pomatomus saltatrix
Centropristis melana
Sciaenops ocellata
Trachinotus falcatus
Chaetodipterus faber
Caranx hippos
Diplodus holbrooki
-101-
-------
Literature Cited
Aldcrdirc, D. F. and J. R. Brett: Some effects of kraft mill effluent on young
Fucific salmon, J. Fish. Res. Bd. Can. 14, 783-795 (1957).
Ann.*!tronfl, N. E., P. N. Storrs, and E. A. Pearson: Development of a gross toxicity
criterion in San Francisco Bay, Proc. 5th Int. Conf. Way. Pol. Res. Ill, 1/15 (1971).
Bechtcl, Timothy J. and B. J. Copeland: Fish species diversity indices as indi-
cators of pollution in Galveston Bay, Texas. Contrib. Mar. Sci., Texas 15
(December), 103-132.
Beck, William M., Jr.: A stream quality survey of the Fenholloway River,
Bureau of Sanitary Engineering, Fla. State Board of Health, Jacksonville,
Florida, February (1954).
Bergcr, W. H. and F. L. Parker: Diversity of planktonic foraminifera in deep-sea
sediments, Science 168, 1345-1347 (1970).
Betts, J. L. and C. G. Wilson: New methods for reducing the toxicity of kraft mill
bleachery wastes to young salmon, J. Fish. Res. Bd. Can. 23(6), 813-824 (1966).
Bocsch, D. F. : Species diversity of marine macrobenthos in the Virginia area,
Chesapeake Science 13(3), 206-212 (1972).
Boescli, D. F. : Classification and community structure of macrobenthos in the
Hampton roads area, Virginia, Mar. Biol. 21, 226-244 (1973).
Brillouin, L.: Science and information theory academic press, 2nd edition. New
York (1962).
Colby, Peter J. and Lloyd L. Smith, Jr.: Survival of Walleye eggs and fry on paper
fiber 6ludge deposits in Rainy River, Minnesota, Trans. Amer. Fish. Soc.
96(3), 278-296 (1967).
Cole, A. E.: Effects of industrial wastes on fish, Sewage Works J. 7, 280-302
(1935).
Conne]1, J. H.: A predator-prey system in the marine intertidal region, 1.
Balnnus glandula and several predatory species of Thais, Ecol. Monogr.
40, 49-78 (1970).
Copeland, B. J.: Effects of industrial waste on the marine environment, J. Water
Pollut. Contr. Fed. 38(6), 1000-1010 (1966).
Copeland, B. J.: Estuarine classification and responses to disturbances, Trans.
Amer. Fish. Soc. 99, 826-835 (1970).
Courtright, Robert C. and Carl E. Bond: Potential toxicity of kraft mill
effluent after oceanic discharge, Progr. Fish Cult. 31(4), 207-212 (1969).
Dahlberg, M. D. and E. P. Odum: Annual cycles of species occurrence abundance
and diversity in Georgia estuarine fish populations, Amer. Midi. Natur.
83(2), 382-392 (1970).
-102-
-------
Das, B. S., S. G. Reid, J. L. Betts and K. Patrick: Tetra-chloro-o-benzoquinone
as a component in bleached kraft chlorination effluent toxic to young salmon,
J. Fish. Res. Bd. Can. 26, 3055-3067 (1969).
Dayton, P. K.: Competition, disturbance, and community organization: the provision
and subsequent utilization of space in a rocky intertidal community, Ecol.
Monogr. 41(4), 351-389 (1971).
Dayton, P. K. and R. R. llessler: Role of biological disturbance in maintaining
diversity in the deep sea, Deep-sea Res. 19, 199-208 (1972).
Fujiyn, Ma sum: Effects of kraft pulp mill wastes on fish, J. Water Pollut.
Contr. Fed. 33, 968 (1961).
FuJJya, Masaru: Physiological estimation on the effects of pollutants upon
aquatic organisms, Advances in Water Pollution Res. vol. 3, E. A. Pearson,
ed., Water Pollution Control Federation, Washington, D.C., p. 315-331 (1965).
Goto, Miyako and Tadashi Goto: Fish fauna in River Nagaragawa and its change for
the last several years in relation to river pollution, Jap. J. Ecol. 21(516),
254-270 (1972).
Grande, M.: Water pollution studies in the River Otra, Norway. Effects of pulp
and paper mill wastes on fish, Int. J. Air Water Pollut. 8,77-88 (1964).
Hairston, N. G.: Species abundance and community organization, Ecology 40, 404-416
(1959).
Hansen, D.J.: Avoidance of pesticides by untrained sheepshead minnows, Trans.
Amcr. Fish. Soc. 98(3), 426-429 (1969).
Hansen, D.J., E. Matthews, S. L. Nail, and D. P. Dumas: Avoidance or pesticides
by untrained mosquitofish, Gambusia affinis, Bull. Environ. Contam. Toxicol.
8(1), 46-51 (1972).
Heck, Kenneth L. Jr.: The impact of pulp mill effluents on species assemblages
of epibcnthic invertebrates in Apalachee Bay, Florida, M.S. thesis, Florida
State University, Tallahassee (1973).
Hicks, D. B. and J. W. DeWitt: Effects of dissolved oxygen on kraft mill effluent
toxicity, Water Res. 5, 693-701 (1971).
Holland, G. A., J. E. Lasater, E. D. Neumann and W. E. Eldridge: Toxic effects
of organic and inorganic pollutants on young salmon and trout, Wash. State
Dep. Fish Res. Bull. 5, 264 p. (1960).
Holland, J. S., Maciolek, Nancy J., and C. H. Oppenheimer: Galveston Bay benthic
community structure as an indicator of water quality, Contrib. Mar. Sci., U.
of Texas Mar. Sci. Inst. 17, 169-188 (1973).
Hooks, Theresa Ann: An analysis and comparison of the benthic invertebrate com-
munities in the Fenholloway and Econfina estuaries of Apalachee Bay, Florida,
M.S. thesis, Florida State University, Tallahassee (1973).
Hopkins, A. E., P. S. Galtstoff and H. S. McMillan: Effects of pulp mill pollution
on oysters (Qstrea lurida), Bull. Bur. Fish. 47, 125-186 (1931).
-103-
-------
Horn, II. S.: Measures of "overlap" in comparative ecological studies, Amer.
Natur. 100, 419-424 (1966).
Hurlbert, Stuart H.: The nonconcept of species diversity: A critique and
alternative parameters, Ecology 52(4), 577-586 (1971).
Tsiuc , S. : Behavior of fish exposed to toxic substances, Advances in water
pollution research; Proc. 2nd Int. Conf., Tokyo (1965).
Ito, T., and K. Kuwada: Aquatic communities and polluted streams with indust-
rial mining waste: 1. Effect of the paper mill waste on the benthic
invertebrates, Annu. Rep. Noto. Mar. Lab. 4, 33-43 (1964).
Jones, B. F., C. E. Warren, C. E. Bond, and P. Doudoroff: Avoidance reactions
of salinomd fishes to pulp mill effluents, Sewage Ind. Wastes 28(11),
1403-1413 (1956).
Katz, M., and A. R. Gaufin: The effects of sewage pollution on the fish pop-
ulation of a midwestern stream, Trans. Amer. Fish. Soc. 82: 156-165 (1953).
Kocnig, Christopher C., Livingston, Robert J., and Claude R. Cripe: Blue crab
mortality: interaction of temperature and DDT residues, Bull. Environ.
Contam. Toxicol (1974).
Lewis, I'rank Graham, III: Avoidance reactions of two species of marine fishes
to kraft pulp mill effluent, M.S. thesis, Florida State University, Tallahassee
(1974).
Livingston, Robert J., Kenneth Heck, Jr. and Theresa Hooks: The ecological Impact
of pulp mill effluents on aquatic flora and faunna of north Florida: Comparison
of a polluted drainage system (Fenholloway) with an unpolluted one (Econfina),
Unpublished report to the Coastal Coordinating Council, Florida; U.S.A. (1972).
Livingston, Robert J., Iverson, Richard L., Estabrook, Robert H., Keys, Vernon E.
and John Taylor, Jr.: Major features of the Apalachicola Bay System: Psy-
siography, Biota, and Resource management, Fla. Acad. Sci. (in press) (1974)
Lloyd, M. : Weighting individuals by reproductive value in calculating species
diversity, Amer. Natur. 98, 190-192 (1964).
MacArthur, R. H.: Environmental factors affecting bird species diversity, Amer.
Natur. 98, 387-397 (1964).
McErlcan, A. J., O'Connor, S. G., Mihursky, J. A. and C. I. Gibson: Abundance,
diversity, and seasonal patterns of estuarine fish populations, Estuarine
and Coastal Marine Science 1(1), 19-36 (1973).
McNaughton, S. J.: Structure and function in California grasslands, Ecology 49,
962-972 (1968).
McNulty, J. Kneeland: Ecological effects of sewage pollution in Biscayne Bay,
Florida: Sediments and the distribution of benthic and fouling macro-
organisms, Bull. Mar. Sci. Gulf Carib. 11(3), 394-447 (1961).
Margalef, D. R.: Information theory in ecology, Gen. Syst. 3, 36-71 (1958).
-104-
-------
Morista, M.: Measuring of interspecific association and similarity between
communities, Mem. Fac. Sci. Kyushu Univ. (3) 3, 65-80 (1959).
O'Conno.r, J. S.: The benthic macrofauna of Moriches Bay, New York, Biol. Bull.
142(1), 84-102 (1972).
Oiilaug, T. 0.: Effects of stabilized and unstabilized waste sulphit0 liquor
on the Olympia oyster, Ostrea lurida, Trans. Amer. Microsc. Soc. 67, 163-182
(1948).
Odum, H. T. , Cuzon DuRest, R.P., Beyers, R. J. and C. Allbaugh: Diurnal metabolism,
total phosphorous, ohle anomaly, and zooplankton diversity of abnormal marine
ecosystems of Texas, Publ. Inst. Mar. Sci. Univ. Texas 9, 404-453 (1963).
Odum, W. E.: Insidious alteration of the estuarine environment, Trans. Amer. Fish.
Soc. 99, 836-847 (1970).
Paine, R. T.: Food web complexity and species diversity, Amer. Natur. 100, 65-75
(1966).
Patrick, Ruth: A study of the numbers and kinds of species found in rivers in
eastern United States, Proc. Acad. Natur. Sci. Philadelphia (1961).
Patten, B. C. : Species diversity in net phytoplankton of Raritan Bay, J. Mar.
Res. 20, 57-75 (1962).
Pearson, T. H.: Effect of pulp mill effluent on fauna of Sea Loch, Marine Poll.
Bull. 1(6), 92-94 (1970).
Pearson, T. H.: The benthic ecology Loch Linnhe and Loch Eil, a sea loch system
on the west coast of Scotland. III. The effect on benthic fauna of the
introduction of pulp mill effluents, J. Exp. Mar. Biol, and Ecol. 6, 211-253
(1971).
Peer, D. L.: Effects of kraft mill effluents on a marine benthic community, Wat.
Air, Soil Pol. 1, 359-364 (1972).
Pielou, E. C.: Species-diversity and pattern-diversity in the study of ecological
succession, J. Theor. Biol. 10, 370-383 (1966).
Pielou, E. C.: The measurement of diversity in different types of biological
collections, J. Theor. Biol. 13, 131-144 (1966).
Pielou, E. C.: The use of information theory in the study of the diversity of
biological populations, Proc. of Fifth Berkeley Symp. on Mathematical
Statistics and Probability, Statistical Lab., Univ. of Calif. 6/21/65-7/18/65
and 12/27/65-1/7/66 (1966).
Quammen, M. L., P. A. LaRock & J. A. Calder: Environmental effects of pulp mill
wastes, Pres. at the Belle W. Baruch Symposium in Marine Science on Estuarine
Microbial Ecology, June 8, 1971 (Columbia, S.C.), 29 pp. (1971).
Reish, D. J.: The uses of marine invertebrates as indicators of water quality,
Waste disposal in the Marine Environments (E. A. Pearson, Ed.) Proc. 1st.
Intern. Conf. 92-103. Pergamon: New York (1960).
-105-
-------
Sanders, II. I..: Marine bcnthic diversity: A comparative study, Amer. Natur.
202„ 243-282 (1968).
Savillc, Thorndike: A study of estuarine pollution problems on a small unpolluted
estuary and a small polluted estuary in Florida. Engineering progress at the
University of Florida Vo. 20(8), (Bulletin series #125) pp. 1-202 (1966).
Servizi, J., E. Stone, and R. Gordon: Toxicity and treatment of bleach kraft mill
waste, Int. Pac. Sal. Fish. Comm., 13, 1-34 (1966).
Shannon, C. L., and W. Weaver: The Mathematical Theory of Communication, Univ. of
Illinois Press, Urbana (1963).
Shumway, Dean L. and George G. Chadwick: Influence of kraft mill effluent on the
flavor of salmon flesh, Water Res. 5(11), 997-1003 (1971).
Simberloff, Daniel: Properties of the rarefaction diversity measurement, Amer.
Natur. 106(949), 414-418 (1972).
Simpson, E. H.: Measurement of Diversity, Nature 163, 688 (1949).
Slobodkin, L. B.: Experimental populations of Hydrida., J. Anim. Ecoc. 33
(Suppl.), 131-148 (1964).
Smith, L. L. and R. H. Kramer: Survival of walleye eggs in relation to wood fibers
and Sphaerotilus natans in the Rainy River, Minnesota, Trans. Amer. Fish. Soc.
92, 220-234 (1963^
Smith, L. L., R. H. Kramer and J. C. McLeod: Effects of pulpwood fibers on fathead
minnows and walleye fingerlings, J. Water Pollut. Contr. Fed., 37(1), 130-140
(1965).
Sprnguc, J. B.: Avoidance of copper-zinc solutions by your salmon in the laboratory,
J. Water Pollut. Contr. Fed. 36(8), 990-1004 (1964).
Sprague, J. B.: Avoidance reactions of Rainbow Trout to zinc sulphate solutions,
Water Res. 2, 367-372 (1968).
Sprague, J. B., P. F. Elson and R. L. Saunders: Sublethal copper-zinc pollution
in a salmon river- a field and laboratory study, Advances in Water Pollution
Research, Proc. 2nd Int. Conf., Tokyo, 1, 61-75 (1965).
Sprague, J. B. and D. W. McLeese: Toxicity of kraft pulp mill effluent for larval
and adult lobsters, and juvenile salmon, Water Res. 2, 753-760 (1968a).
Sprague, J. B. and D. W. McLeese: Different toxic mechanisms in kraft pulp mill
effluent for larval and adult lobsters, and juvenile salmon, Water Res.
2, 761-765 (1968b).
Sprague, John B. and Donald E. Drury: Avoidance reactions of salmonid fish to
representative pollutants, Adv. in Water Poll. Res., 4th Inter. Conf. on Water
Poll. Res., Pergamon Press New York, 169-186 (1969).
-106-
-------
Sumincrfelt, R. C. ami W. M. Lewis: Repulsion of green sunfish by certain chemicals
J. Water Pollut. Contr. Fed., 39(12), 2030-2038 (1967).
Swartz, Richard C., DcBen, Waldemar A. and Andrew J. McErlean: Comparison of speci
diversity and faunal homogeneity indices as criteria of change in biological
communities, Proc. L.P.A. Symposium on Monitoring the Marine Environment,
Seattle, Wash. (1973).
Taylor, CJydc C.: Nature of variability in trawl catches, Fish. Bull. Fish Wild.
Service 54, #83, 145-166 (1953).
Trnmer, Elliot J. and Philip M. Rogers: Diversity and longitudinal zonation in
fish populations of two streams entering a metropolitan area, Amer. Midi.
Natur. 366-374 (1973).
Tsaj, Chu-Fa: Effects of chlorinated sewage effluents on fishes in upper
Patuxcnt River, Maryland, Ches. Sci. 9: 83-93 (1968).
Van Horn, W. M., J. B. Anderson and M. Katz: The effect on kraft pulp mill wastes
on some aquatic organisms, Trans. Amer. Fish. Soc. 79, 55-62 (1949).
Waldcn, C. C., T. E. Howard, G. C. Froud: A quantitative assay of the minimum
concentration of kraft mill effluents which affect fish respiration, Water
Res. 4(1), 61-69 (1970).
Waldichuk, M.: Some water pollution problems connected with the disposal of
pulp mill wastes. Can. Fish. Cult. 31, 3-34 (1962).
Webb, P. W. and J. R. Brett: The effects of sublethal concentrations of whole
bleached kraft-mill effluent on the growth and food conversion efficiency
of undcryearling sockeyc salmon, J. Fish. Res. Bd. Can., 29(11) 11/72
1555-1563 (1972).
Weiser, Conrad Earnest: A comparison of the avifauna between an unpolluted marsh
and a polluted marsh on the Gulf coast of Florida, M.S. Thesis, Florida
State University; Tallahassee (1973).
Wildish, D. J., Carson, W. G., Carson, W. V. and J. H. Hull: Effects of a
neutral-sulphite, pulp effluent on some chemical and biological parameters
in the L'Etang Inlet, New Brunswick. L'Etang Inlet Survey I., Manuscript
Report Scries #1177, Fisheries Research Board of Canada, May (1972).
Wiley, A. J., B. F. Lueck, R. H. Scott, and T. L. Wismiewski: Cooperative
state-industry stream studies - lower Fox River, Wisconsin, Ind. Wastes
29(1) , 76-87 (1957).
Wilhm, J. L. and T. C. Dorris: Species diversity of benthic macroinvertebrates
in a stream receiving domestic and oil refinery effluents, Amer. Midi. Natur.
76(2), 427-449 (1966).
Wilhm, J L., and T. C. Dorris: Biological parameters for water quality criteria,
Bioscience 18, 477-481 (1968).
-107-
-------
Wilson, John N. : Effect of kraft mill wastes on stream bottom fauna, Sewage
Incl. Wastes 25(10): 1210-1218 (1953).
Wilson, R. C. H.: Acute toxicity of spent sulphite liquor to Atlantic salmon
(Salino salar), J. Fish. Res. Bd. Can. 29, #8, 1225-1228 (Aug. 1972).
Woelkc, C. E.• Effects of sulphite waste liquor on the normal development of
the Pacific oyster (Crassostrea gigas), Research Bull. State of Washington
Fish. Dept. 6, 149-161 (1960).
Woelkc, Charles E.: Bioassays of pulp mill wastes with oysters, Biological
Problems In water Pollution - Third Seminar 8/13-17, U.S. Public Health
Service (1962).
Woelkc, C. I;.: Application of shellfish bioassay results to the Puget Sound
pulp mill pollution problem, Northwest Sci. 42, 125-133 (1968).
Ziebcll, C. D., R. E. Pine, A. D. Mills and R. K. Cunningham: Field toxicity
studies and juvenile salmon distribution in Port Angeles Harbor, Washington,
J. Water Pollut. Contr. Fed. (1970).
Zimmerman, Michael S.: A comparison of the benthic macrophytes of a polluted
drainage system (Fenholloway River) with an unpolluted drainage system
(Hconfma River), M. S. Thesis, Florida State University, Tallahassee
(1974).
-108-
-------
SUMMARY
1. A 2 year study was carried out to determine the effects of pulp mill effluents on
the fishes of a river, estuary, and coastal system in north Florida (U.S.A.) A
comparison was made with the fauna of a nearby uncontaminated river and coastal
sysLem.
2. Repetitive trawl tows were analyzed by computerized methods utilizing various
diversity indices. This allowed maximal use of the collecting method as well as
n basis for comparative analysis of the data.
3. The Fenholloway River and estuary (and associated marsh areas) had high levels of
color and turbidity and low dissolved oxygen. Offshore areas were characterized
by varying levels of PME; in general, inshore stations proximal to the Fenholloway
River mouth were especially affected. Runoff patterns, wind, and tidal phenomena
were major factors in the distribution of the effluents. Elevated turbidity and
color and seasonally reduced benthic D.O. levels were characteristic of the
affected offshore stations.
4. Except for occasional topminnows that were swept into the river from surrounding
areas, and some fishes that moved periodically up the salt wedge of the estuary,
there were no permanent fish assemblages found in the Fenholloway River, estuary,
or adjacent marsh areas during the study period.
5. Offshore areas that were affected by PME had significantly lower numbers of fishes
(N) than control areas. Although the cumulative number of species caught in each
system was approximately the same, the average number of species (S) taken per
month at affected (Fenholloway) stations was significantly below that of control
stations. Differences in N and S between affected and control stations were
minimal during winter months.
-109-
-------
Qualitative differences were noted in the fish distributions of affected
~nd unaffected areas. The inshore stations of the Fenholloway drainage
system were dominated by the bay anchovy (Anchoa mitchilli). With the
exception of one Fenholloway station, areas stressed by PME had lower
dominance values.
Species diversity and species richness indices were lower at only the
most grossly stressed stations, and the usefulness of such parameters as
indicators of pollution was questioned. Species diversity was considered to
be important when used in conjunction with various other parameters.
Areas on either side of the stressed offshore system showed dramatic (periodic)
increases in numbers of individuals, numbers of species, and various species
diversity indices. This contrasted with relatively constant evenness values
throughout the entire area of study.
-110-
-------
List of Tables
Comparison of different parameters relative to significant levels
of variation (p
-------
List of Figures
Fig. 1: Chart showing portions of Apalachee Bay receiving drainage
from the Econfina River, the Fenholloway River, and Spring
Warrior Creek.
Fig. 2: Seasonal variations of mean sea water temperatures in the
2 offshore study areas over a period of 25 months.
Fig. 3: Color and salinity data at all stations over a 26 month period.
Fig. 4: Turbidity data at all stations over a 25 month period.
Fig. 5: Comparison of the dissolved oxygen levels in the Fenholloway
and Econfina Rivers at different seasons of the year with
attention to salinity relationships.
Fig. 6: Comparison of the dissolved oxygen (7, saturation) in benthic
areas of offshore stations over a 24 hour period (August, 1971).
Samples were taken simultaneously in the 2 study areas for
comparison.
Fig. 7: Two year comparison of the number of individuals (N) and species (S)
of fishes taken from 4 "inner" stations (F9, F10, E7, E8)
Fig. 8: Comparison of repetitive (7 trawl) fish data (N & S) on all
regular offshore Fenholloway and Econfina stations (11 month
totals for "outer" stations; 12 month totals for "inner" stations)
Fig. 9: Dominance-diversity curves of paired stations (Econfina-
Fenholloway) summed over period of 1 year.
Fig. 10: Scaled species diversity (ENS) and species richness (SR) of
paired stations (Econfina-Fenholloway)
Fig. 11: Number of individuals (N) and number of species (S) for
transect stations
Fig. 12: Brillouin (H) and Shannon-Weaver ( H'*) species diversity,
evenness (J1), and scaled expected number of species (E)
for transect stations
Fig. 13: A. Trellis diagram based on C X values of summed transect data
B. Trellis diagram based on cXvalues of summed data for
regular stations.
Fig. 14: Composite diagram showing strong relationships (CA >0.9 ) of
sampling stations of the Econfina-Fenholloway-Spring Warrior
Creek area.
-112-
-------
Fig. 1: Chart chowing portions of Apalachee Bay receiving drainage
from the Econfina River, the Fenholloway River, and Spring
Warrior Creek.
GJ
tn
T?1
TIO
ru
no-
-------
Fig. 2: Seasonal variations of mean sea water temperatures in the
2 offshore study areas over a period of 25 months.
tfY
fO
-. Econflno
—F*nhoilo«oy
MOJFUAUJJAS
TIME IN MONTHS (1971-73)
F '< U
t
-------
Fig. 3: Colo* and salinity data at all stations over a 26 month period.
O
v
CJ1
I
o
w
r|
1
•
tconri^A
t
•
»
x..a/^Wv_ -A
»
^ /^A-^ _ A
-AAA
9
\
1
• A ' A \
\y \
° • •
f\
J-
'V^A
" • * - ' 5
"* • h
4
j
a .
t
J ^\/~\ _ f-
"¦""7 •
£
•. /_
%
P'
at
o
mi
©
V
o
©
v
TIMt IK MOUTH S (HII 75)
!¦ « B| |i I 90
70
a-"Vv
V »
: *0
,v
\ A
v \
TO
--v. / V \^..
A-^a_W
to
_ J,
n/*n''v v\ A
A\- V-wA
cw
IOC
>•
r
«0 «i
<
- t'\ A
V,.N
••¦ s~''\
v^A/t
lO*"
if #V«.v
- Sx.\
•a
—r-
10
14 A * t
c ,#*s » 1/«' \/
#s / •- * • • \ / *%
to
-v -"n
*" v.; ».•¦
•^AW,4/
/ \ ¦—I | % / * * » *
o / ¦» v / V' - ;.
/ • S ' \ .
J- • •/ j
k- A.A ~'\l
u
/«
1
H .0
1
1 A
TIME IN MON T H S { 197173)
-------
Pig. A: Turbidity data at all stations over a 25 conch period.
' CAM 0 UOtAY
CCO«nMA
A/
\
TT
~ IMC IN WONT Nl (|»TI*7»)
Wr
't«MOUO**T
•• i £ C C k 11 H A
•4
-¦ijil'i'i"
O
a
at
\
14
TIME IN MOHYMB (l»7l*7))
-------
Fig. 5: Comparison of the dissolved oxygen levels in che Fenholloway
and Econfina Rivers at different seasons of the year with
attention to salinity relationships.
ECONFINA
a .
a <6
5-
>-
x
O to
-J
o
CO
(0 CM
c/
! S 3o
•TATION
CO •
£
tx
a *
z
Ul tf)
<9
>-
X
o t
o
Ui
>
J
o
V)
0)
10
/
| 1-.|—| I — .— •
2 3 4 5 5a 5b 6 9 10 16
April, 1971
balinit y
• m r f o C • » ~m
bottom •
/
y'm
4 4a 5 5 7 6 9
February, 1971
//
V
7
o -
N Q-
a.
O >¦
- H
Z
<0
i 3 3a 4 4a 5 6 7 5 9 2 3 3a 4 4a 5 6 7 8 9
February, 1972 July,1972
FE NHOLLOWAY
2 3 4 5 5a 5b 6 9 10 16 2 3 4 5 5a 5b 6 9 10 16
Jonuory, 1972 June,1972
OISSOIVEP OXYGEN
• vrfot« » ¦¦ i »
bottom • ¦ •
-117-
-------
Flf,. 6: Couip.irison of the tlis^nlvcd oxygen (7, saturation) in benthic
nrcns of offshore stations over a 24 hour period (August, 1971).
Samples were taken simultaneously in the 2 study areas £or
comparison.
DISSOLVED OXYGEN ( % S AT U R AT ION )
-118-
-------
Yig.
Two year comparison o£
of fishes l.ikcn from 4
Che number of. individuals (N) and species (S)
"inner" staLions (F9, F10, E7, E8)
NUMBER OF INDIVIDUALS (N)
ci
NUMBER OF SPECIES (S)
o ~
Ul
01
ro
O
£w.. \p
\ '""¦•K--..
• O \
-119-
-------
I'ig. 8:
Comparison of rcpcliLivc (7 Lrnwl) fish data (N & S) on all
regular offshore Fcnholloway and Kconfina stations (11 month
totals for "outer" sLations; 12 month totals for "inner" slat
ions)
number of individuals
f ¦ f . f
f , ?
»•
a
II II
» -
V
¦c
T ff
Ik 0
» «
ii ii a
II -
o
II II _
ii * ©
U tl
n ii _
*• <0 o>
«* s
o
X
H
X
w
(0
M
II II
r t
¦
J
II II
M » — L
! i
¦ n _
g a —
•» r w
• i
m a
o i —
!
w r w
!
II I! «»
H tt W
« M
f X
NUMBER OF
SPECIES
-120-
-------
Fig. 9: Douiinnnce-diversit y curves of paired stations (Econfina-
Fcnholloway) summed over period of 1 year.
1000_
100
10.
1.
1000
100.
10
1.
1000.
> 100J
o
10.
1.
1000
100.
10.
F11
• • ••**•••!
E 12
F12
E 11
10
20 30
4^0
F14
iwtwmi
10 20 30
E 13
F IS
«»i»mn
«••••••••
E 7
F9
• • tWMKI..!
E 8
F10
'Sl.
1 ¦ ¦ ¦ Tf| ^ * *
E 9 10 20 30 40f 16 J0 2'0 3!0
E TOTAL•••
F. TOTAL-
Vs
V,
«
I,
~l 1 1 1 1 1 1 1 1 1 1 1—"if ~r i
10 20 30 40 10 20 30 40 10 20 30 40 50 60 70 80
SPECIES
-121-
-------
Fig. 10: Scaled species diversity (ENS) and species richness (SR) of
paired stations (Econfina-Fenholloway)
-122-
-------
F1B. 11: Number of indivuiu.-iU cms
transect stations -nd number of species (S) for
VI
¦j
<
o
a
a
x
u.
O
o
z
ISO-
iOO-
N *-
s •
no t« tj3 tsi t2o rn Fio fis tit n? 122 tss
TRANSECT STATIONS
-123-
-------
12: Brillouin (H)
evenness (J1)
for transect
and Shannon-Weaver (
, and scaled expected
stat ions
H ) species diversity,
number of species (E)
1.0
0.5
J U LY
2.5-
2.0-
X
X
1.0-
0 5-
-1.0
-0.5
25. NOVEMBER
10
2.0-
x
6
1.5-
x
1.0-
-2
0.5-
-1.0
-05
2.j. APRIL
-10
2.0-
X
15-
X
-4
10-
0.5-
>3
77
19
F10 F1S
17
10 Fll
E10 E8 Ell
11
)' *
h' wmooa
(SCALED)C ©• ®
STATION
-124-
-------
Fig. 13: A. Trellis diagram based on C X values of summed transect data
R, Trellis diagram based on cXvalues of summed data for
regular stations.
EIO
Et
EI3
T2I
120
Fll
FIO
FI5
TI7
TI9
T22
T23
EIOE8 EI3 T21120 Fll FIO FI5 TI7TI9T22T23
.94
\
.67
.85
.74
AO
.34
.51
.54
.56
.61
.57
mm
.sc
.95
.90
.15
.44
.67
.6£
.74
.70
.70
\
.93
\
.94
.24
.44
.CG
.84
.86
.94
.71
•< .tcJiyi
.94
.21
.23
.80
.76
.78
.88
.69
\l
.24
.48
.80
.84
.86
.90
.80
.40
.38
.56
.49
.39
31
.68
.71
.59
.51
.89
.90
.94
.61
\.
.97
r\
.94
.79
.94
.77
.76
E8 E9 EIOEII EI2 EI3F9 FIO Fll FI2FI4FI5 FI6
B
> 90 ~
.80-89
.70-.79 H
.30—.69 | [
<.50
E
: xUQ:
-125-
-------
Fig. 14: Composite diagram showing strong relationships (cX > 0.9 ) of
sampling stations of the Econfina-Fenholloway-Spring Warrior
Creek area.
-126-
-------
A Review of Clustering Techniques
with Emphasis on Benthic Ecology.
By
John D. Walker
Western Interstate Council for Higher Education Intern
Coastal Pollution Branch
U.S. Environmental Protection Agency
Marine Science Center
Newport, Oregon 97365
August 1974
Committee Members:
R.
C.
Swartz
D.
T.
Marti n
W.
A.
Deben
D.
J.
Baumgartner
-127-
-------
ABSTRACT
Numerical clustering techniques are
Hierarchical, Divisive Hierarchical
The use of these various techniques
reviewed.
reviewed including Agglomerative
and Non-Hierarchical strategies,
in Marine Benthic Ecology is also
-128-
-------
INTRODUCTION
Numerical grouping techniques have been used in a variety of scientific
fields, especially numerical taxonomy. In the past decade, ecologists
have employed these methods to elucidate cormiunity structure. Because
of this diversity of interests, these techniques have been referred to
as taxometric methods (Sneath and Sokal, 1973), classification methods
(Pielou, 1969), and clustering techniques (Williams, 1971). Clustering
techniques seems the most general term because it does not imply a
specific use.
In benthic studies, the raw data consists of a number of species taken
from several collection sites. Clustering techniques analyze this raw
data giving an investigator some idea of the similarities within the
entire group. Benthic studies using clustering techniques usually
analyze the data two ways: collection sites are considered as the
individuals with the species from the sites acting as attributes
resulting in site groups; or species are considered as individuals
with collection sites as attributes resulting in species groups.
Because of the ability of clustering techniques to sort out community
types, it has become of interest in pollution studies, as was done
recently by Stephenson et al_. (1974) and Roback ert al_. (1969).
Because of pollutional stresses, benthic cormiunities affected by
pollution will be sorted from community types in healthy areas,
thus facilitating detection of pollution stresses and management
of pollution abatement efforts.
Clustering techniques can be divided into three main subgroups, two
of which are hierarchical and the third non-hierarchical or reticulate.
-129-
-------
The two hierarchical strategies are referred to as agglomerative and
divisive. Both strategies define a method of producing clusters of
similar individuals from the entire population. Agglomerative
hierarchies start with the individuals and fuse groups up to the
entire population while divisive hierarchies start with the population
and divide it down to the individuals. Non-hierarchical strategies
do not define a route between groups but optimize the structure of the
group by making it as homogeneous as possible.
Of the two hierarchical strategies, agglomerative seems to have more
inherent problems than divisive; however, in marine ecology, the
agglomerative hierarchical strategies seem to be more popular. The
two main problems with agglomerative strategies are: (1) since a user's
interest is normally concentrated in the higher levels of the hierarchy
the fusion process must run from the individual to the entire population.
For a population of N individuals the fusions necessary equal (N-l)
(Williams, 1971). For a large N, computation time can be prohibitive.
(2) There Is a tendency for minor misclassification because the fusion
starts at the individual level where chance anomalies are more likely
to occur.
Divisive strategies are not as sensitive to these two problems. Since
fission starts at the population, the chance anomalous behavior of the
individuals is more likely overlooked. Also, starting with the entire
population a "stopping rule" can easily be programmed to halt fission
at whatever level the investigators interest lies, hence computation
time can be reduced considerably. Most divisive strategies are based
on MONOTHETIC division (i.e., based on a single attribute, which in the
case of benthic studies would be presence-absence) and selecting the
right attribute requires considerable insight since it must divide the
population into two groups as unlike as possible. Polythetic divisive
strategies which divide the population based on more than one attribute
-130-
-------
(such as % similarity or dissimilarity) do exist; however, they
require a much greater computational time than agglomerative
strategies of comparable size.
In this paper, the two hierarchical strategies will be discussed in
detail including benthic studies where these methods were employed.
Non-hierarchical strategies will also be discussed although they
have not received the attention that hierarchical systems have.
AGGLOMERATIVE HIERARCHICAL STRATEGIES (AHS)
Interindividua! Measures To initiate the fusion in an agglomerative
strategy, there needs to be some measure for comparing individuals.
They are concerned with numerical definition of likeness and have been
referred to as interindividua! measures by Williams (1971). There have
been large numbers of these measures proposed (reviewed by Bergen, 1971;
Goodman and Kruskal, 1959; and Sokal and Sneath, 1963), all of which
fall into three main classes (Williams, 1971). These are (a) Manhattan
metric of the basic form zlX^ - X2J|, (b) Euclidean distance
^'Xlj " x2jt2)15 (In both cases for site SrouP classification and
X2j would be the number of individuals of species (j) in sites 1 and 2),
and (c) various forms of information statistics usually using the
formulation of Shannon (the specific formulation varies, examples of
which will be given later).
Williams (1971) outlined several decisions that must be made prior to
clustering. Double-zero matches (which in the case of site classification
would mean the species in question was absent from both sites) which are
quite common in species rich marine surveys are of particular interest.
If double-zero matches are counted toward likeness, the interindividual
measure is symetrical; if they are not, then it is asymetrical (Williams,
1971). Both symetrical and asymetrical models are available so the
-131-
-------
choice is left to the individual user. Since double-zero matches are
not a specific problem of AHS, further discussion of it as well as
transformation of attributes and standardization will be discussed in
the last section of the paper.
Individual/Group and Group/Group Measure Every AHS begins with the
calculation of all interindividual measures. As soon as fusion begins,
individual/group and group/group measures are needed. It is desirable
that the individual/group and group/group measures are similar to the
interindividual measure for then both the interindividual measure and
individual/group measure can be considered a special case of the group/
group measure. Some of the measures contain what is described by Lance
and Williams (1967b) as Confoinatorial solutions. In such cases, the
individual/group and group/group measures can be calculated directly
from a matrix of interindividual measures. This has the advantage
that once the interindividual measures are calculated, the raw data
is no longer needed for subsequent calculations.
Lance and Williams (1967b) have further shown that the majority of the
combinatorial solutions can be encompassed within a single generalized
linear model: dhk = a, dh, + o.j dhj ~ 8 + y|dM - dhj|. In this
model, d^, d^. and d^j are all dissimilarity type measures for the
individuals of groups h, i and j. Groups i and j are fusing to form
group k for which the equation calculates the dissimilarity between
h and k (d^). Values for , a j, 3 and y vary depending on the
strategy.
The following strategies are mostly those given by Lance and Williams
(1967b). Those with combinatorial solutions which can be calculated
by the linear model above will have the appropriate values for o^., otj,
3 and y.
-132-
-------
Nearest-neighbor. The distance between the two groups is
defined as the distance between their closest individual,
one in each group (a. = a. = +h, 8 = 0, y = -h).
' J
Furthest-neighbor. It is the opposite of Nearest-neighbor
in that the distance between two groups is defined as that
between the most remote pair of elements, one in each group
(a.| = otj = +H» 3 = 0, y = +H).
Centroid. On fusion the fused individuals are replaced by
the co-ordinates of its centroid which is the sum or mean
sum of the individuals forming the group.
i. Squared Euclidean Distance. If the co-ordinate of
the centroid of (i) is X; then the centroid of the
new fused group (K) will be (N.X. + N.X./NJ. Then
II J J K
the difference measure dhk for groups (h) and (K)
will be d^ = {Xh - (N.X^ + Nj/N^}. The centroids
of groups (i), (j) and (h) are denoted by X,. . hx
vi» J»
and N., N., and N„ are the number of individuals in
1 J l\
groups (i), (j) and (K). The strategy for this
measure is
«i = N./N^, = Nj/Nk, 3 = - and y = 0.
ii. Correlation Coefficient. For qualitative data, the
Pearson coefficient is usually used and for
quantitative data the product-moment coefficient is
used. This is = combinatorial, but requires two
equations for calculation so it cannot be calculated
by the single linear equation.
iii. Non-Metric Coefficient. For binary data, the
complement of Czekanowski coefficient is normal.
For quantitative data where X,. and X . are the
number of individuals in the J'ih spec?4s in site
one and two for site classification
*1X1.1 ' X2.jl
could be used (Lance and Williams, 1967a).
-133-
-------
d. Median. This strategy is similar to squared Euclidean
measure except N. is put equal to N. which results in
' V
ot. = a. = +h, 3 = -!a, Y = 0.
' J
e. Group-average. This defines the intergroup distance as
the mean of all the between-group interindividual
distances (Stephenson et al_., 1972). If there are N.
members in group i and N. members in group j so that in
the fused group K there are members, then
= H/N,,, a. = iyNK, 3 = Y = 0.
f. Incremental sum of squares. The numerical model is
Euclidean; the decision-function is the increase on
fusion of the sum of the squares of the distance
between the individuals and their group centroids
(Williams, Clifford and Lance, 1971). Its
combinatorial properties are not known.
g. The flexible strategy of Lance and Williams (1967b) is
derived from the linear model (a- + a. + 3 = 1, a. =
a., 3 F co, y = 0). It is compatible Tor Euclidean
distance and derives the flexibility from its space
distorting properties which will be discussed below.
h. The information statistic strategy is an (i, j, k)
measure derived from the information content before
and after fusion by the relationship A I,, , , ^ =
I. - I. - I.. It is a non-combinatorial4trategy.
k i j 3,7
In clustering strategies where an element placed in a group has no
effect on its original position in the space, it is said to be space
conserving. This property is possessed by any classificatory strategy
which uses Euclidean distance between group centroids as its inter-
group measure and for AHS of Euclidean systems using centroid fusion
strategy. In contrast, the more intensely clustering AIIS, the groups
appear to recede from each other as they grow. These strategies are
called Space Dilating. The receding of groups is group size dependent.
This dependence on group size may take one of two forms. It may be
asymptotic, so that once the group has attained a modest size further
accretions make little difference; or it may be indefinite, so that
-134-
-------
every accretion makes the group substantially more remote and
therefore substantially more difficult to join (Williams, 1971).
Individual/group measures may differ in this respect from the
group/group measures. For example, in flexible sorting both measures
are asymptotic; in the incremental sum of squares the individual/group
measure is asymptotic, the group/group measure indefinite; for the
commonest information statistic strategy both are indefinite (Williams
et aK, 1971).
Choice of Strategies In agglomerative hierarchical clustering
strategies, one has to make a choice of the interindividual measure
to employ and then the fusion strategy used in building the hierarchy.
The choice of interindividual measures is not difficult because it
usually is dependent on the nature of the data (Williams, 1971).
Highly skewed binary data obtained from presence-and-absence records
would usually require Shannon-type information statistics since it is
insensitive to skewness and handles such attributes without difficulty
(Williams et al_., 1971). Euclidean measures which are unduly sensitive
to the rare occurences handle data defined by a small number of
continuous variables with no strong outliers (Williams, 1971). In '
contrast, most information statistics are inefficient in dealing with
continuous variable (Williams et al_., 1971). If the data is non-
negative with few zeros, but with an occasional extreme outlier, the
Canberra metric is indicated (Williams, 1971). In cases where the
data have no striking pecularities, the choice of measure is of
relatively little importance, with mixed data it has been observed
that information statistics and incremental sum of squares tend to
produce nearly identical classifications (Williams, 1971).
The choice of the fusion strategy is much more difficult than the
choice of an interindividual measure because there are few hard
guidelines to follow. An investigator may have no idea as to the
-135-
-------
clustering character of his data or he may wish for greater accuracy
at the expense of clearly separate clusters. In such a case, the
weakly clustering strategies are the most desirable within the AHS.
The appropriate strategies would be nearest-neighbor, group average
or centroid strategies. Williams et aK (1971), feel that the group
average strategy has some disadvantages in that the system is
indefinite. If an attribute is highly skewed with a small number of
outlying values, the variance will rise abruptly on fusion with the
first of these tending to leave further outlyers stringing along
unfused. It also measures identical groups with zero variance so
a group of identical individuals acts as an individual, thus very
similar individuals forming a group may be vulnerable late in the
analysis to capture of an outlying element better placed elsewhere.
The more intensely clustering programs are more appropriate for those
workers whose data is either largely continuous and he wishes to chop
them into groups as efficiently as possible, or his data may be highly
heterogeneous requiring a space dilating strategy. As mentioned earlier,
intensely clustering strategies share the comnon property of increasing
the 1ntergroup similarity as group size increases resulting in possible
non-conformist groups whose members share only the property that they
are unlike everything else including themselves (Lance and Williams,
1967b). A way of relieving this problem is to develop a method for
reallocating individuals after the hierarchy is formed to better
allocate individuals from a non-conformist group or individuals that
were misclassified due to the inherent tendency in AHS for slight
misclassification.
The space dilating strategies discussed by Williams et al_. (1971)
include information statistic, incremental sum of squares and
flexibility. Information statistic as mentioned earlier has a strong
tendency to produce a non-conformist group, the position of a group
in the hierarchy will be strongly dependent on the size of the group
and no effective reallocation procedure is known.
-136-
-------
Incremental sum of squares has little tendency to produce non-
conformist groups although the position of a group in the hierarchy
is strongly dependent on the size of the group. Reallocation has
been shown to be simple and effective. Flexible strategy does not
tend to form the non-conformist group, nor have a marked size
dependence on groups entering the hierarchy.
Use of AHS in Benthic Community Analysis As mentioned earlier,
agglomerative hierarchical strategies have been the most popular
for studying community structure of the macrobenthos. Stephenson
and Field have both published several papers where these methods
have been used with good results. Although there are no rigid
guidelines on which clustering strategy should be used for an
individual set of data, the work that has been done to date seems
to Indicate that one or two strategies are most useful. The habitats
that have been analyzed with clustering techniques have been quite
diverse ranging from tropical benthos, intertidal, semi-tropical and
temporate estuarine, yet the same strategies seem to prevail.
The interindividual measure used most commonly is the complement of
Czekanowski or Bray-Curtis similarity index C2 = 2W/(A+B) where A is
the sum of the measures of all species in one sample, B is the similar
sum for the second sample and W is the sum of the lesser measures of
each species for the two samples being compared. This measure was
used by Field, 1968 and 1971; Field and McFarland, 1968; Stephenson
and Williams, 1971; Stephenson, Williams and Cook, 1972; and
Stephenson, et al_., 1974. Another interindividual measure was used
1 N | X* • - X«. |
several times, the Canberra Metric Dissimilarity d, 9 = ^ E > ' ^ .
' M i 1*11 *2i;
where X^ and X2i are the importance values of the ith attribute of the
two individuals. That measure was used by Stephenson, Williams and
Cook (1972) and Boesch (1973).
-137-
-------
The fusion strategies that found the greatest success were the group
average and the flexible with 3 = -0.25. Field (1968, 1971) used
group average only as did Stephenson e_t aj_- (1974). Boesch (1973),
Stephenson, Williams and Lance (1970), Stephenson, Williams and Cook
(1972) used both flexible with 3 = -0.25 and group average. Boesch
found almost identical results between flexible and group average for
the collection site analysis but felt the flexible sorting was much
superior to group average for the species analysis. Stephenson and
Williams (1971) tried two different information statistic clustering
strategies and in both cases found fragmentation of aberrantly rich
sites and haphazard nonsense fusion between poorer sites. They
resorted to the flexible sorting using 3 = -0.25 with satisfactory
results.
DIVISIVE HIERARCHICAL STRATEGIES (DHS)
In describing the DHS and AHS earlier, it seems that AHS are fraught
with problems that are not encountered with DHS. A natural question
would be, why haven't the AHS died out completely in preference for
the DHS? The reason is that most practicing DHS programs are based
' on monothetic division. In contrast, all AHS are polythetic systems
based on similarity or dissimilarity. Although the monothetic
classifications are simple and fast, they are easy to misclassify.
Suppose two groups, X and Y, to be separated monothetically on an
attribute possessed by X and lacking in Y. Also suppose individual
B possesses the attribute which clusters it with X but more closely
resembles the members of Y. At a later stage B will be separated
from the main division sterming from X but it will not be able to
gain access to the Y side. Because of this, it is characteristic of
divisive monothetic systems to produce unduly large numbers of
fragmentary groups (Williams, 1971). The polythetic divisive
strategies would not do this since it measures overall similarity.
They would therefore seem the ideal hierarchical strategy, however,
-138-
-------
existing programs require a very small population size and the new
program of Wallace.and Boulton (1968) are untested on field problems
(Wi11i ams, 1971).
Monothetic Divisive Two strategies which are based on monothetic
division are Association Analysis and Information Analysis. In
Association Analysis, the critical attribute is determined by an
Index of Association (I^). The Index of Association in this case
is a group measure which determines which should be the next group
to be divided or when division should stop. In association analysis
of sampling sites, Ift would be calculated for every possible pair of
species and the values of 1^ entered in an association matrix. The
elements of the association matrix are then summed by columns or rows
and the critical species to divide the sites on is that having the
greatest value of E Ift (Pielou, 1969). The index value IA has been
defined in three ways by Williams and Lambert (1959) as x2» X2
corrected and /x /N«
The information statistic measure proposed by Lance and Williams
(1968) calculates the information content, I, of a population as
I = SN Log N - § (a. Log a, + (N-a. Log (N-a.)} where N site^are
3=1 J J J J
defined by S species such that the jth species is present in aj
individuals. If a population of sampling sites is divided into two
groups (g) and (h), then the information fall would be defined as AI
(gh, i) = I.. - I - 1^. The population is divided at each stage by
the attributes for which AI is maximum.
Polythetic Divisive As mentioned earlier, divisive-polythetic
strategies are more complex and much more time consuming to calculate
than divisive monothetic methods. The methods of Edwards and Cavalli-
Sforza (1965) have the attributes represented by a single point in a
S-dimensional space. The group measure consists of the sum of the
-139-
-------
squared distances of all points to the group centroid (Pielou, 1969).
The division of a group is based on reducing the within-group sum of
squares to a minimum which results in a maximum between-group sum of
squares. The method is summarized by Pielou (1969). A population
represented by 41 points would take 54,000 years to complete using a
computer with access time of 5y seconds (Gower, 1967). As a result,
this method has not found much popularity with benthic ecologists.
A more complicated divisive-polythetic strategy is that proposed by
Wallace and Boulton (1968) and Boulton and Wallace (1970) based on
information measure. Their program instead of basing group comparison
on a measure of similarity or dissimilarity is based on the probability
of a member of a group possessing certain measured attributes. The
mathematics involved in the strategy are far too complex to discuss in
this paper. No mention of computation time is made and since there is
no record of anyone using this program in a large study, it remains to
be seen whether it would be appropriate for benthic studies.
Use of DHS in Benthic Community Analysis The use of divisive
programs in benthic studies has been rather sparse. Because of its
greater computational speed over agglomerative and divisive polythetic
strategies, divisive monothetic strategies are appropriate with studies
containing a great number of stations and/or species.
Moore (1973) employed both association analysis and information
analysis in analyzing the fauna of kelp holdfasts. For association
2
analysis, he used /x /N as the association parameter. The program
p
was "flagged" when no individual x exceeded 3.84. Although there
was some slight misclassification with both strategies, he felt the
information analysis gave the best results, but that both strategies
were quite satisfactory considering the size of the stucjy.
-140-
-------
Stephenson al_. (1970) used information analysis for their divisive
strategy. Again, as with Moore, the number of species was exceedingly
large, as well as the number of collection sites, so the monothetic
approach was the only practical method. After site groups were
established using the divisive classification, an agglomerative/
polythetic method was used for the species classification using the
site groups, not the individual sites, as attributes.
N0N-HIERARCHICAL STRATEGIES
Williams (1971) says of non-hierarchical systems: "for applications
in which homogenity of groups is of prime importance, the non-
hierarchical strategies are in principle very attractive; unhappily
their current state of development lags far behind that of their
hierarchical counterparts which at their best are more flexible,
provide a wider range of facilities are numerically better
understood, and are computationally faster." Perhaps because of
these disadvantages, the non-hierarchical systems have received
little interest in benthic studies.
The non-hierarchical strategies are divided in two types by Williams
(1971). The first type is serially optimized, that is, a group is
defined and removed from the population. The remaining individuals
are examined and a second group removed; this continues until the
entire population is accounted for. The second type is simultaneously
optimized where the population is partitioned some way into groups
and these are optimized by a repetitive process.
Initiation of a non-hierarchical strategy begins with the calculation
if interindividual measures much like the AHS. Later allocation to
groups is primarily an individual/group measure. To determine if an
individual should be added to a particular group, there have been
several sorting strategies proposed most of which Lance and Williams
-141-
-------
(1967c) feel are mathematically or computationally unsatisfactory.
Of the serially optimized methods discussed, only that of Goodall
(1966) was considered mathematically sound.
Goodall's method is based on a probabilistic similarity index used
for interindividual measures. Clusters are then formed of all
individuals whose maximum dissimilarity is less than a maximum
based on a specified significance level. Of the various clusters
formed, the largest is removed and the process continued with the
residue. The repeated computation of the similarity matrices which
is an integral part of the system makes computation time quite lengthy
and therefore not suitable for problems of a very large size.
The simultaneously optimized strategies reviewed by Lance and
Williams do not suffer from the mathematical inconsistencies of
the serial group. Their favorite is MacQueen's (1966) K-mean system.
The population is arbitrarily partitioned into K groups and the mean
of their Euclidean distance is calculated. As new members are added
to the group, the new means are calculated. As a result, the K mean
shifts as allocation proceeds. If two group means come closer than a
predetermined value, they are fused reducing the number of K groups.
Since an upper limit is fixed on allocating an individual to a group,
it is possible that an individual cannot be fused so a new nucleus is
formed and K increases.
Lance and Williams do not discuss the recurrent group analysis of
Fager and McGowan (1963), but it would be considered a serially
optimized strategy. The strategy based on interspecific affinities
using J/^aNb)55 as the interindividual measure. In this index, J
is the number of joint occurrences, N is the total number of
a
occurrences of species A, Nb is the total number of occurrences of
species B. A choice of a significant affinity is chosen and the
species are allocated into the largest groups where all species
-142-
-------
have affinities equal to or greater than the chosen significance
level. The procedure was originally used by Fager and McGowan for
zooplankton, but in more recent time, has been used in benthic surveys
by Jones (1969), Lie and Kelly (1970), and Boesch (1971).
ADDITIONAL CONSIDERATIONS
Transformations If the range in the numbers of species is very wide,
as it often is in marine surveys, the transformation In (Nj+1) may be
used, where N. is the number of individuals of the jth species. If
J
the counts are small, they may be distributed as Poisson variates
where transformation will normally be the distribution (Stephenson
and Wi11iams, 1971).
Untransformed data was used in the study of Stephenson and Williams
(1971) and Stephenson et (1970). In the study of the LA Bight,
Stephenson et al. (1974) used the cube root transformation.
Standardization of the interindividual dissimilarity measures is
also carried out by some workers. Stephenson et^ al_. (1974) used
unstandardized measures for the site-groups but standardized by
totals for the species-groups. Boesch (1973) used double
standardization for site and species groupings.
Double Zero Matches The occurrence of double-zero matches in marine
studies where species numbers are usually quite large is connton. They
are normally dropped in benthic studies but problems can arise whether
they are included towards similarity or not. If they are included and
there are many of them, the analysis will be dominated by groups having
only zero in common (Williams, 1971). If they are ignored and there
are some impoverished sites or very heterogeneous data, a hodge-podge
of unfusable sites may result or the fusing with any group that contains
a species in cornion (Fields, 1969).
-143-
-------
A more difficult problem ecologically is how to handle 0-1 up to
perhaps 0-5 matches. Using the Canberra metric dissimilarity
coefficient, a 1-0 match has the same dissimilarity as 100-0 match,
liocsch (1973) to offset this, suggests substitution of a small constant
for zero which he set at 0.001. A number so small has little effect
even at the 1-0 level and can't be considered much improvement.
Further improvement in this area will hopefully be forthcoming.
Chaining This is a problem in the hierarchical strategies only.
Some weakly clustering or space conserving strategies appear on
formation of a group to move nearer to some or all the elements.
This is not true of all space conserving strategies and has been
labelled space-contracting by Lance and Williams (1967b). With
such a strategy, the chance of an individual element adding to a
pre-existing group is greater than an individual forming the nucleus
of a new group. This system is said to be chaining, that is, the
tendency of a group to add single individuals or groups much smaller
than itself, rather than fusion with groups of comparable size
(Williams, Lambert and Lance, 1966). The flexible sorting
strategies used in AHS with positive values for 8 have strong
tendencies for chaining. It is felt by Williams e_t al_. (1966)
that the more symmetrical the hierarchy, the better the clustering
techniques.
Choice of Strategies An investigator deciding to use cluster
techniques on his data has wide range of strategies from which to
make his choice. Oftentimes, the decision is made for him due to
pragmatic reasons such as program availability, computer time, etc.
If he has several programs to choose from, his first decision would
probably be between a hierarchical and non-hierarchical strategy
depending on whether he wanted the cluster as homogeneous and error-
free as possible or was willing to sacrifice minor misclassification
-144-
-------
for more rigid, well-separated groups. Lance and Williams (1967c)
suggest a double analysis using the hierarchical system initially,
at a subdivision within the hierarchy the data would be analyzed
non-hierarchically using a simultaneously optimized strategy such
as the K-mean. Once the investigator has made the decision between
non-hierarchical or hierarchical, he is mostly on his own.
If the user decides on a hierarchical strategy, the size of the study
(i.e., the number of collection sites or number of species collected)
may be taken into account. In the case of Moore (1973) where 387
species and 72 sampling sites were involved in the analysis, he felt
from a computational standpoint that the divisive monothetic strategy
was the only practical choice. Stephenson et al. (1970) analyzed 387
sites and 355 species by first eliminating the rare species (those
occurring six times or less). After reduction, they were left with
358 sites and 51 species which were analyzed using a divisive
monothetic strategy. The resulting site-groups, eight in all, were
then used with an agglomerative strategy for the species-groups.
The methods used in the examples above might be appropriate if the
study were very large. However, if the study is not as large as
those mentioned above and computer time is not a primary consideration,
one of the agglomerative strategies would be the most advisable.
-145-
-------
REFERENCES
Bergan, T. 1971. Survey of numerical techniques for grouping.
Bacteriological Reviews 35:379-389.
Boesch, D. F. 1971. Distribution and structure of benthic
communities in a gradient estuary. Ph.D. Dissertation,
College of William and Mary, Virginia.
Boesch, D. F. 1973. Classification and community structure of
macrobenthos in the Hampton Roads Area, Virginia. Marine
Biology 21:226-244.
Boulton, D. M. and C. S. Wallace. 1970. A program for numerical
classification. The Computer J. 13:63-69.
Edwards, A. W. F. and L. L. Cavalli-Sforz. 1965. A method for
cluster analysis. Biometrics 21:362-375.
Fager, E. W. and J. A. McGowan. 1963. Zooplankton species-groups
in the North Pacific. Science 140:453-460.
Field, J. G. 1968. The use of numerical methods to determine
benthic distribution patterns from dredgings in False Bay.
Tras, Royal Soc. South Africa 39:183-200.
Field, J. G. 1971. A numerical analysis of changes in the
soft-bottom fauna along a transect across False Bay, South
Africa. J. Experimental Marine Biology and Ecology 7:215-253.
Field, J. G. and G. McFarlane. 1968. Numerical methods in marine
ecology. 1. A quantitative "similarity" analysis of rocky
shore samples in False Bay, South Africa. Zoologica Africana
3:119-137.
Field, J. G. 1969. The use of the information statistic in the
numerical classification of heterogeneous systems. J. Ecology
57:565-569.
-146-
-------
Goodal1, D. W. 1966. A similarity index based on probability.
Biometrics 22:882-907.
Goodman, L. A. and W. H. Kruskal. 1959. Measures of association
for cross-classification. 0. American Statistical Association
54:123-163.
Gower, J. C. 1967. A comparison of some methods of cluster
analysis. Biometrics 23:623-637.
Jones, G, F. 1969. The benthic macrofauna of the mainland shelf
of Southern California. Allan Hancock Monograph in Marine
Biology 4. 219 pp.
Lance, G. N. and W. T. Williams. 1967a. Computer programs for
hierarchical polythetic classification. The Computer J. 9:60-64.
Lance, G. N. and W. T. Williams. 1967b. A general theory of
classification sorting strategies. I. Hierarchical Systems.
The Computer J. 9:373-380.
Lance, G. N. and W. T. Williams. 1967c. A general theory of
classificatory sorting strategies. II. Clustering Systems.
The Computer J. 10:271-277.
Lance. G, N. and W. T. Williams. 1968. Note on a new information
statistic classificatory program. The Computer J. 11:195.
Lie, U. and J. C. Kelley. 1970. Benthic in fauna communities off
the coast of Washington and in Puget Sound: Identification
and distribution of the communities. J. Fisheries Research
Board of Canada 27:621-651.
MacQueen, J. B. 1966. Some methods for classification and analysis
of multivariate observations. Western Management Science
Institute, Univ. of Calif., Working Paper #96.
Moore, P. G. 1973. The kelp fauna of Northeast Britain. II.
Multivariate classification: Turbidity as an ecological
factor. J. Experimental Marine Biology and Ecology 13:127-163.
Pielou, E. C. 1969. An introduction to mathematical ecology.
Wiley-Interscience. New York. 286 pp.
-147-
-------
Roback, S. S., J. Cairns and R. L. Kaesler. 1969. Cluster analysis
of occurrence and distribution of insect species in a portion
of the Potomac River. Hydrobiologia 34:484-502.
Sneath, P. H. A. and R. R. Sokal. 1973. Numerical taxonomy.
W. H. Freeman and Co. San Francisco. 573 pp.
Sokal, R. R. and P. H. A. Sneath. 1963. Principles of numerical
taxonomy. W. H. Freeman and Co. San Francisco. 359 pp.
Stephenson, W., W. T. Williams and G. N. Lance. 1970. The
macrobenthos of Moreton Bay. Ecological Monographs 40:459-494.
Stephenson, W. and W. T. Williams. 1971. A study of the benthos
of soft bottoms, Sek Harbour, New Guinea, using numerical
analysis. Australian J. Marine and Freshwater Research
22:11-34.
Stephenson, W., W. T. Williams and S. D. Cook. 1972. Computer
analysis of Petersen's original data on bottom conmunities.
Ecological Monographs 42:387-415.
Stephenson, W., R. W. Smith, T. S. Sarason, C. R. Greene and D. A.
Hotchkiss. 1974. Soft bottom benthos from an area of heavy
waste discharge, Palos Verdes, California. I. Hierarchical
classification of data from 87 sites. Submitted to J.
Experimental Biology and Marine Ecology.
Wallace, C. S. and D. M. Boulton. 1968. An information measure
for classification. The Computer J. 11:185-194.
Williams, W. T. 1971. Principles of clustering. In_: Annual
Review of Ecology and Systematics. 2:303-326.
Williams, W. T. and J. M. Lambert. 1961. Multivariate methods
in plant ecology. III. Inverse association analysis. J.
Ecology 49:717-729.
Williams, W. T., J. M. Lambert and G. N. Lance. 1966. Multivariate
methods in plant ecology. V. Similarity analyses and
information analysis. J. Ecology 54:427-446.
-148-
------- |