,. TECHNICAL
in aquatic *****
environment research 324
Part 1 1 - Biological assessment of
marine pollution with particular
reference to benthos
by
John S. Gray
Institute for Marine Biology and Limnology
University of Oslo, Norway
Alasdair D. Mclntyre
Marine Laboratory
Aberdeen, United Kingdom
Jose Stirn
Fisheries Science and Technology Department
Sultan Qaboos University, Muscat, Oman
FOOD
Prepared in cooperation with:
/^^ ^
> ^ United Nations Environment Programme
Mediterranean Action Plan
lliau, I99O
The designations employed and the presentation of material in this
publication do not imply the expression of any opinion whatsoever on
the part of the Food and Agriculture Organization of the United
Nations concerning the legal status of any country, territory, city or
area or of its authorities, or concerning the delimitation of its frontiers
or boundaries
M-42
ISBN 92-5-103136-3
All rights reserved No part of this publication may be reproduced, stored in a
retrieval system, or transmitted in any form or by any means, electronic, mechani-
cal, photocopying or otherwise, without the prior permission of the copyright owner
Applications for such permission, with a statement of the purpose and extent of the
reproduction, should be addressed to the Director, Publications Division, Food and
Agriculture Organization of the United Nations, Viale delle Terme di Caracalla,
00 100 Rome, Italy
FAO 1992
PREPARATION OF THIS DOCUMENT
FAO participates in the implementation of the Long-term Programme for
Pollution Monitoring and Research in the Mediterranean (MED POL) - Phase II, which
is coordinated by the United Nations Environment Programme. In the framework of
the MED POL programme, Mediterranean Institutions undertake research to study the
ecosystem modifications in areas influenced by pollution* During Phase I of the
programme a manual was developed entitled "Manual of Methods in Aquatic
Environment Research, Part 8 - Ecological assessment of pollution effects' 1 (FAO
Fish. Tech. Pap (209), 1981), which aimed to contribute to the identification of the
effects on marine life of pollutants from different sources. The manual was
prepared by Professor J. Stirn.
The FAO/UNEP Meeting on the effects of pollution on marine ecosystems
(Blanes, Spain, 7-11 October 1985) recommended that the above manual should be
updated in order to maintain its usefulness as a guide to appropriate research
techniques* Another conclusion of the meeting was that there is an urgent need
for training in methods of data analysis. FAO/UNEP organised such training
courses in cooperation with the Group of Experts on the Effects of Pollution of
the Intergovernmental Oceanographic Commission (IOC/GEEP). Training workshops on
the statistical treatment and interpretation of marine community data have so far
been organised in Piran, Yugoslavia (1988), Athens, Greece (1989), split,
Yugoslavia (1990) and Alexandria, Egypt (1991).
The present manual was prepared by Professor A*D. Mclntyre and
Professor J.S. Gray and draws to a great extent on material from Part-8 of this
series mentioned above and from the course lecture material. For this reason
Professor J. Stirn has been retained in the authors.
Final editing and compilation was done by the staff of the FAO Fishery
Resources and Environment Division, particularly Mr. G.P. Gabriel ides. Ms V.
Papapanagiotou was responsible for the typing.
The views expressed in the manual are those of the authors and do not
necessarily represent the views of either FAO or UNEP.
DEFINITION OP MARINE POLLUTION
Pollution of the marine environment means: "The introduction by man,
directly or indirectly, of substances or energy into the marine
environment (including estuaries) which results in such deleterious
effects as harm to living resources, hazards to human health,
hindrance to marine activities including fishing, impairment of
quality for use of sea water and reduction of amenities".
IMO/FAO/Unesco/WMO/WHO/IAEA/UN/UNEP Joint Group of Experts on the
Scientific Aspects of Marine Pollution (GESAMP)
Cover photo: Sampling of benthos in the Mediterranean using a Smith-Mclntyre grab.
Photograph by Dr V.A. Catsiki, National Centre for Marine Research,
Athens, Greece
Distribution:
FAO Fisheries Department
FAO Regional Fisheries Officers
Mailing list Inland Pollution
Marine Pollution
Gray, J.S.; Mclntyre, A.D.; Stirn, J.
Manual of methods in aquatic environment research. Part 11. Bio-
logical assessment of marine pollution with particular reference
to benthos.
FAQ Fisheries Technical Paper. No. 324. Rome, FAO. 1991. 49p.
SUMMARY
Chemical analysis, although valuable and necessary, does not
provide all the information required in pollution assessments.
Biological studies are of particular value in permitting a realistic
assessment of pollution and they cover a wide range of possibilities.
The present manual makes only cursory reference to the techniques
used to study the sublethal toxic effects at the "individual" level
of organisation or below since it is devoted to biological studies at
community level and especially to the use of benthos. It describes
how a benthic sampling programme should be designed so that the data
collected can be best interpreted and evaluated. Information is
provided for the collection and treatment of the samples as well as
for the analysis of the data using statistical methods and computer
software. Multivariate analysis techniques include hierarchical
clustering, multi-dimensional scaling (MDS) ordination and principal
components analysis (PCA).
TABLE OF CONTENTS
Page
1 . INTRODUCTION 1
2. THE BIOLOGICAL APPROACH TO MARINE POLLUTION STUDIES 2
3. BIOLOGICAL TECHNIQUES USED IN MARINE POLLUTION ASSESSMENT 3
3.1 Biological studies below the level of the individual
organism 3
3.2 Biological studies at the level of the individual
organism 4
4. BIOLOGICAL STUDIES AT POPULATION LEVEL AND ABOVE 7
4.1 The use of plankton 7
5. THE USE OF BENTHOS 8
5.1 Choice of sample site 8
5.2 Acquisition of data and preliminary investigations 8
5.3 Design of a benthic sampling programme 9
5.3.1 Size and number of replicate samples
per site 10
5.3.2 Temporal sampling frequencies 12
5.4 Sampling and processing methods 12
5.4.1 Intertidal sampling 13
5.4.2 Sub tidal sampling: Ships and shipboard
equipment 13
5.4.3 Subtidal sampling on hard bottoms 14
5.4.4 Subtidal sampling on soft bottoms 14
5.4.4.1 Qualitative and semi-
quantitative sampling 14
5.4.4.2 Quantitative sampling 15
5.4.5 Other techniques 17
5.4.5.1 Diving 17
5.4.5.2 Photography 19
5.4.6 Treatment of samples 20
5.4.6.1 Macrofauna 21
5.4.6.2 Meiofauna 23
5.5 Analyses of results 26
5.5.1 Discrimination between sites 27
5.5.2 Univariate methods determining levels
of disturbance 28
5.5.2.1 Between site comparisons 28
5.5.2.2 Determination of disturbance
at individual sites 34
5.5.2.3 Correlation with anthropogenic
inputs 34
5.5.2.4 Analyses of higher taxa only 36
6. REFERENCES 39
1. INTRODUCTION
There are many definitions of pollution, but the GESAMP formulation is
widely accepted and constitutes part of the protocol of several international
agreements and conventions. It states: "Pollution means the introduction by man,
directly or indirectly, of substances or energy into the marine environment
(including estuaries) resulting in such deleterious effects as harm to living
resources, hazards to human health, hindrance to marine activities including
fishing, impairment of quality for use of seavater, and reduction of amenities' 1
(Pravdic, 1981).
This definition first indicates that marine pollution arises from substances
added to the sea by man. Although these substances cover the whole chemical
spectrum they can be classified into a relatively small number of categories.
These include:
nutrients
sewage
oil
metals
synthetic organic compounds
radionuclides
plastic litter
particulates
The definition secondly recognizes that pollution implies adverse effects
on the environment. Nutrients, mainly nitrogen and phosphorus, whether derived
from urban waste waters, from industrial discharges, from agricultural run-off or
from natural weathering of the land, act as biostimulants, causing eutrophication
- an enhancement of the growth of seaweeds and phytoplankton. This can lead to
the development of unusual plankton blooms which may or may not be toxic but which
on decay use up oxygen from the water with adverse consequences for fish and
invertebrates. Sewage is another source of nutrients but in addition it
contributes large amounts of organic matter which also causes deoxygenation.
These effects of nutrients are found particularly in sheltered areas where water
exchange with the open sea is restricted, and where there is considerable
urbanization or industrialization, or where large rivers draining agricultural
land reach the coast. The inner Adriatic is an obvious example, and also the Bay
of Izmir and the Gulf of Lions. As well as affecting the plankton, eutrophication
can alter the structure of benthic communities. Sewage presents an additional
public health threat in that it carries pathogenic organisms that can cause
disease in human beings from contamination of seafood and beaches. Oil enters the
sea via operational discharges from ships and offshore installations, from
accidents and from various coastal effluents. In large amounts it smothers
habitats and organisms, and fresh oil has toxic components. It forms slicks on
the sea surface damaging seabirds and marine mammals, while weathered oil washed
ashore as tar balls reduces beach amenity. Although there have so far been no
major oil spills in the Mediterranean, it is a busy shipping zone with many tanker
routes, and drilling for oil and gas is currently conducted by eight countries in
the region.
Metals are natural components of seawater and sediments, and as such are
harmless to marine life, but they can build up to high concentrations as a result
of Man's activities, as in mine tailings or industrial effluents, and may then
represent a risk to human consumers of seafood. It is not always easy to
partition the residues in animals between natural and anthropogenic sources. For
example in some countries of the Mediterranean Sea (Algeria, Italy, Spain, Turkey
and Yugoslavia) weathering of natural cinnabar deposits is thought to contribute
to the relatively high level of mercury measured in some marine organisms, but in
certain areas industrial effluents are also relevant.
Synthetic organic compounds, particularly pesticides (e.g. DDT) and certain
industrial chemicals (e.g. PCBs) are now widely distributed in the environment.
Being fat-soluble, persistent and largely non-biodeeradable they accumulate in
sediments and in the lipids of organisms. Their build-up in top predators,
particularly marine mammals and birds, causes damage and their presence in seafood
can make it unacceptable for human consumption. Some organometals such as TBT can
be toxic at low concentrations, and chronic effects, including imposex in the
gastropod Nucella. can be induced at very low concentrations.
Although radioactivity tends to arouse public concern, radiation from
artificial radioactive substances is extremely low, reaching even the level of
natural background in only a few localities. Radionuclides enter the sea from both
natural and anthropogenic sources, but at concentrations which do not pose a
threat to marine organisms. The main input from Man is in wastes and the largest
quantity is derived from nuclear fuel reprocessing. This is not done on an
industrial scale in the Mediterranean so most releases into that sea are from the
25 nuclear power stations which operate in four countries on the northern
Mediterranean coast - France, Italy, Spain and Yugoslavia. Since most of those
power stations are located alongside freshwater, their effluents are transported
to the sea via river systems.
Plastic litter is a relatively new category of marine contaminant, but is
causing concern. The synthetic materials now widely used for fishing nets,
packaging straps and containers are buoyant and persistent. Discarded nets
continue to trap animals by 'ghost' fishing, straps and rings encircle mammals,
birds and fish, and plastic materials of all kinds accumulate on beaches.
A detailed evaluation of these pollutants and their effects on a global
basis has been made by GESAMP (1990) and a related exercise involved a study of
conditions in each of UNEP's Regional Seas areas. In the Mediterranean report it
is stated that all confined or semi -confined localities adjacent to large urban
centres appear to be in a state of progressive build-up of pollution as a result
of anthropogenic release, while eutrophication is recorded as a significant
problem.
Since pollution is related to the introduction of substances to the sea,
a first approach to its measurement can be made by the chemical analyses of
contaminants in the water, in the sediments and in the tissues of organisms. Such
data on chemical concentrations can indicate actual or potential problems,
especially if linked with information relating concentrations to effects, but by
themselves chemical measurements do not constitute assessment of pollution. That
assessment can properly be made only by observing in the field biological effects
on the biota.
2. THE BIOLOGICAL APPROACH TO MARINE POLLUTION STUDIES
As indicated above, chemical analysis, although valuable and necessary, does
not provide all the information required in pollution assessments. Indeed, it is
not the concentrations of contaminants per se which are of concern, but rather the
effects of these concentrations on organisms and on human health, and it is only
by documenting these effects that the true significance of the chemical data can
be defined. It is for this reason that biological effects studies are an
essential component of any pollution assessment programme, and this was early
recognized by FAO in the series of publications under the general heading "Manual
of Methods in Aquatic Environment Research", some of which are referred to, as
appropriate , below .
It is important, however, to clarify what is implied here by 'effect'. In
a sense the accumulation of chemical residues in the tissues of organisms is an
effect of the inputs and also the concentrations of contaminants in the water and
sediments. But in terms of the present discussion a biological effect is more
specific. It occurs when the organisms can be shown to react in some way to the
contaminant. Knowledge of biological effects therefore provides information on
the impact of the contaminant on the biota and since that impact may cause changes
in population and communities of organisms, the acceptability of this can be
considered, and the ultimate evaluation of the pollution can be made. Quite apart
from providing this ultimate assessment, the use of the biological approach has
certain practical advantages. One constraint of chemical analyses is that they
are usually directed towards specific contaminants and will therefore omit any
that are not on the suspect list, possibly missing a key toxic chemical. Studies
of biological effects, on the other hand, pick up and integrate responses to the
totality of chemicals in the water or sediments and so provide a comprehensive
picture.
Biological studies are thus of particular value in permitting a realistic
assessment of pollution, but they also assist in other ways. Since many animals,
particularly filter- feeders, accumulate chemicals from the water and sediments
into their bodies (Portmann, 1976), they act as sentinels in providing an early
warning of potential problems. Also, in the bioassay mode, test organisms can be
used in experimental situations to appraise environmental quality (Reish and
Oshida, 1986). Finally, at an even more finely focused level, organisms are used
in what is usually referred to as toxicity tests, which include among other
things, screening tests, tests to develop water quality standards, and legal
tests, which are usually designed to determine LCSOs (see for example,
UNEP/FAO/IAEA, 1987a, b and c) .
Biological effects studies proper cover a wide range of possibilities. An
early workshop in this field (Mclntyre and Pearce, 1980) identified some 50
relevant techniques, ranging through biochemistry, physiology, pathobiology,
behaviour, genetics and finally population and community studies. These aspects
are referred to in the following sections of this manual, but particular emphasis
is given to the study of ecological effects.
3. BIOLOGICAL TECHNIQUES USED IN MARINE POLLUTION ASSESSMENT
This section considers briefly the approaches available and in current use
to assess marine pollution. It focuses on techniques applied at and below the
level of individual organisms, leaving treatment of the higher levels of
organization (populations and communities) to sections 4 and 5.
3.1 Biological studies below the level of the individual organism
A stressor acting on an organism will produce a reaction and it should be
possible to detect the effects at cellular and biochemical levels before these
effects become obvious at the level of whole-animal physiological processes. A
wide range of techniques has been proposed (Uthe et al. . 1980; Moore and Lowe,
1985), or which some are listed below:
Blood chemistry studies, using blood serum assay techniques.
Adenylate energy charge determination, a measure of the metabolic energy
available to an organism from the adenine nucleotide pool.
Metallothionein, which is the metal -binding protein induced in some
organisms by the presence of heavy metals in the water.
Cytochrome P-450 system, which is induced by certain organic contaminants.
Lysosomal fragility, a study of the destabilisation of lysosomes
(membrane -bound sacs containing enzymes, found within the cytoplasm of
animals) by contaminants.
Steroid hormone metabolism studies.
Taurine:glycine measurement, involving a study of changes in the ratios of
some free amino acids.
While these and other techniques have been studied in the laboratory, many
of them would not be appropriate for application in the field on a routine basis,
because of, for example technical problems and difficulties of interpretation (Lee
et al. . 1980). This prompted the Group of Experts on Effects of Pollutants (GEEP,
sponsored by the Intergovernmental Oceanographic Commission of UNESCO, the
International Maritime Organization and the United Nations Environment Programme)
to set up a series of workshops where the most promising techniques were tested
against each other along known pollution gradients. The tests were done 'blind'
in that the research workers involved were not told where the samples had come
from and had to rank the grades of pollution independently (Bayne et al. . 1988),
From this exercise, two reliable techniques have emerged.
First, where the pollution is an organohydrocarbon (PAH, PCB or related
chemical) presence of the pollutant induces activity in the cytochrome P-450
system in both mussels (Mvtilus edulis) and flounder (Platichthvs flesus) . By
measurement of this enzyme it is possible to assess whether or not a likely
pollutant has stressed the organism, (see Stegeman, 1980; Addison, 1984; Stegeman
et al. . 1988; Addison and Edwards, 1988 for methods). However, this approach must
be used with caution, since at certain stages of the fishes' life cycle the P-450
system can be switched off, (eg. when egg formation is taking place) and false
reading can be made. Also some evidence suggests that enzyme system can be induced
only once; if a fish is left to recover fully after exposure to a stressor, a
second exposure may not result in further induction.
Second, heavy metals present in the water induce activity in metal -binding
proteins (metallothioneins) in many marine organisms. Measurement of the presence
of metallothioneins in Mytilus edulis (Viarengo fil^, 1988) has been shown to
be a reliable technique for the measurement of stress induction. Metallotheonein
induction occurs in Polychaeta and some Crustacea and is likely to become a
widespread tool in the future.
Biochemical techniques such as cytochrome (P-450) or metallothionein tests
have the advantage that they are specific to a certain group of chemicals and
indicate what detailed chemical analyses should be made in following up the
investigation.
A number of other tests are available, such as alteration of cell membrane
structure (Moore, 1988), and cytogenetic damage (Perry &L., 1988), but with the
present state of their development, these do not seem as reliable as those
mentioned above.
An approach using genetics is also relevant to studies below the level of
the individual organism. In particular, the cells of fish, especially the eggs
and early developmental stages are very sensitive to chromosome damage arising
from contact with water-borne mutagens, and this can be used to measure the
sublethal effects of sea-surface pollution. Longwell g &L. (1980) provide
details of the techniques and discuss the general approach.
In applying such methods it is important to make proper statistical analyses
and Clarke and Green (1988) provide a useful detailed appraisal of appropriate
statistical procedures.
Using the above techniques, stress can be measured at levels below that of
an individual organism although it can be argued that they do no more than show
that the organism's compensatory mechanism has been induced. For the technique
to be relevant as a measure of biological effects rather than simply a means of
indicating exposure to contamination, some detrimental impact on the growth,
reproduction or survival of the organism must be demonstrated or at least some
link shown with a physiological condition that does have such an impact. In this
context, it is likely that prolonged induction of cytochrome P-450 activity or
metallothionein will lead to effects at the population level, but data are not yet
available to confirm this. Any effect should also be capable of showing a graded
dose/response reaction.
3,2 Biological studies at the level of the individual organism
The biochemical techniques referred to above are responses to more or less
specific stressors. The techniques described in this section are biological
responses that indicate general pollution loadings, or otherwise make particular
use of whole animals.
One obvious effect at the whole-animal level is morphological change, and
a number of such changes linked to contaminants have been documented (ICES, 1978).
These include damage to gill membranes exposed to zinc, gross changes in the liver
caused by pesticides, and various skeletal anomalies involving the gill-rakers,
opercular bones, cranial asymmetries, and vertebral column deformities in fish
exposed to metals and organochlorines . Unfortunately many of the observations are
derived from laboratory experiments in which very high levels of the contaminants
were used. However, there are also field observations suggesting that fish,
especially the younger stages, do exhibit such morphological changes when
associated with pollution hotspots.
One group of morphological effects which is contaminant-specific is related
to tributyl tin (TBT) . This biocide is used as an antifoulant in marine paints,
on shellfish traps and on the structures of fish farms. It causes is shell
thickening in oysters and at very low concentrations it gives rise to imposex
(females developing male characteristics) in some species of gastropods. These
effects are easily detected by eye and are good biological indicators of TBT.
A further possible approach involving morphology is the use of disease-
related changes as an indication of pollution, and this has been advocated
particularly in fish (Sinderman, 1983). To be of value in routine surveys the
effect would need to be one which could be seen readily in the field, without the
need for such detailed laboratory examination as the sectioning of tissues. In
particular, fin erosion, skin ulceration and neoplasms, which occur on the surface
of the fish, are obvious candidates.
Fin erosion is a non-specific disease but many records suggest that it is
associated with degraded coastal or estuarine environments. Two types of fin
erosion can be distinguished. One in demersal fish, probably related to direct
contact with contaminated sediments, occurs mainly on the dorsal and anal fins,
while another in pelagic species is more generalized but predominantly in the
caudal fin. It is important not to confuse mechanical or net-damaged with fin
erosion. True fin erosion is characterized by melanized and darkened tissues
which are not found in net -damaged fish.
Epidermal ulcers have been observed in many fish species (Bucke and
Watermann, 1988) and vary in size from small superficial lesions to large areas
involving skeletal muscle and bone tissue. Ulcers may be caused by microorganisms
such as vibrios or viruses but it has also been suggested that their prevalence
is higher in polluted areas.
Neoplasms, or tumours, have been found in at least 60 marine species from
a variety of habitats (Sinderman et al. . 1980). These may occur on the skin as
epidermal papillomas or carcinomas, or internally as tumours particularly in the
liver. Most species with a higher prevalence of tumours in polluted waters dwell
or feed on the bottom where the concentration of chemicals is usually highest.
The potential of using the prevalence of pathological conditions in marine
animals as a tool in pollution monitoring does seem considerable. The main
problem is that the outward expression of disease depends on highly complex
ecosystem interactions and the separation of natural and anthropogenic impacts is
extremely difficult. Research in this context is at present underway and in the
meantime it is suggested that fish disease indices of pollution should be used
mainly for the initial identification of hotspots and always as only one approach
in a wider suite of studies. As referred to later, the use of fish disease in the
context of populations is even more difficult.
There are three other types of approach involving the use of individual
organisms - bioassays, behavioural studies, and the use of indicator species.
With bioassay tests, a single species is used to monitor the quality of
water or sediment which is suspected of being contaminated. Such tests suitable
for use in the Mediterranean sea have been described by Bellan (1981). One of the
most intensively studied and widely tested bioassays is a physiological index of
stress, scope for growth, in the mussel Mytilus edulis (Bayne, 1980; Bayne and
Worral, 1980; Widdows and Johnson, 1988) which has been shown to be a reliable
indicator of general pollution stress. In this test the energy costs of
respiration and other physiological losses are compared with the food intake and
thereby the physiological well-being of the individual assessed as the potential
energy remaining for growth, the scope for growth. While the test is done on
individual animals by taking replicates good statistical reliability has been
obtained in this assay, (Widdows and Johnson, 1988). It may be argued that the
test does not relate to the population as the response of individuals only is
measured. However, the consequences at the population level of reduced scope for
growth have been well tested, (Widdows, 1985; Koehn and Bayne, 1989) and if scope
for growth is reduced over even relatively short periods of time then both
fecundity and survival of the individual is also reduced and effects will clearly
ensue at the population level.
While a wide range of organisms have been suggested for use as bioassay
tests only a few can be reliably used. In addition to the scope for growth test
in Mytilus. there are tests using oyster larvae (Woelke, 1967, 1968, 1972; Connor,
1972), echinoderm larvae (Kobayashi, 1971), Daphnia (Deneer t al. . 1988) and
Artemia (Abernethy and Mackay, 1986). The references quoted above provide details
of the tests. The organisms are grown in culture under controlled conditions and
performance of the organism is tested in the water to be assayed. For oyster and
echinoderm larvae the percentage metamorphosis is recorded whereas the Daphnia and
Artemia test involves either percentage survival or growth. The tests grade the
assayed water in relation to the control. Since in these tests the responses
measured, i.e. percentage metamorphosis and percentage survival, are at the level
of whole individuals it may be expected that the scope for growth test will be
more sensitive. As yet no comparative experiments have been done.
Another group of organisms which has proved most useful in the bioassay mode
is the hydroids, such as Campanulanias flexuosa. Hydroids are not of commercial
importance, but they are easily handled in experiments because they are sessile,
they are sensitive to stress, and they reproduce asexually so can be cultured as
clones. They thus avoid some of the difficulties presented in working with
larvae. The culture techniques and the operation of tests are described by
Stebbing (1985).
Recently a new approach has been introduced, the sediment bioassay.
Amphipods are known to be particularly sensitive to oil pollution and are
therefore, appropriate to this type of study. The sediment is removed undisturbed
from the natural habitat by a corer, amphipods are placed in the core, and the
time taken to burrow or the percentage survival over 48 hrs is recorded (Chapman
and Lone, 1983; Long and Chapman, 1985). It is possible to rank the sediments in
order of their degree of contamination.
The appeal of behavioural tests is undoubtedly high. It may be expected
that organisms are able to detect a pollutant and initiate an avoidance response
and thereby not be affected. Many techniques have been proposed, involving for
example feeding, ventilation, heart rate, learning and shelter -building (Miller,
1980; Olla et al. . 1980a) . Yet surprisingly few behavioural responses have so far
been adequately studied and/or quantified so that the techniques are not in
general use. However, in concert with other approaches, the use of behaviour does
seem to have much to offer (Olla et al. . 1980b) .
Another use of individual organisms is as indicator species. This involves
recognition that the presence of an individual of a given species in a field
sample may indicate a certain grade of pollution, and a few species have been
proposed for this role. Capitella capitata is a small polyehaete that occurs in
high abundance under conditions of organic enrichment (Reish and Barnard, 1960).
It is found almost world-wide where high levels of organic enrichment occur and
was suggested (Reish, 1970) as a 'universal indicator of organic pollution'.
There are however, problems with the use of such organisms. First, C. capitata
occurs in high numbers under naturally disturbed conditions (Eagle and Rees, 1973;
Grassle and Grassle, 1976a and b) and does not necessarily indicate organic
enrichment. Second, C. capitata has been shown to be a complex of many sibling
species (Grassle and Grassle, 1977) and thus may not in fact be a single
cosmopolitan species. Third, and more seriously in the context of a universal
indicator of organic pollution, it occurs extremely late in the sequence of
organic pollution stress (Pearson and Rosenberg, 1978), and therefore indicates
gross pollution rather than the first states of decline, which is often the object
of pollution monitoring (Gray, 1981). By this stage the sediment has few species
and usually smells strongly of H 2 S, so it is much easier to use this property as
an indicator than to search the sediment for small polychaetes. Thus, except to
indicate extreme polluted conditions, use of an indicator species such as C.
capitata is not recommended.
However, indicator species can be found which are sensitive to subtle
environmental changes (Gray, 1989). These are not likely to be universal, but
rather are specific for given localities.
Before leaving the examination of the use of individual organisms, mention
should be made of the mussel -watch concept. It has long been recognized that
filter- feeding animals, particularly bivalve molluscs, concentrate contaminants
from the water, and Goldberg (1975) proposed that mussels could be used in a
global monitoring programme. In the meantime, this approach is in widespread use
at the regional level, for example, a national mussel watch programme has been
underway since 1975 on both Atlantic and Mediterranean coasts of France, by which
concentrations of several metals and organic contaminants are determined in the
soft parts of mussels and oysters sampled on a quarterly basis (Claisse, 1989).
It should be emphasized that this type of programme uses the animals merely as
integrators of contaminants which are subsequently determined by chemical
analysis. However, if some of the biological effects techniques discussed above
could be built into the exercise, then the value of mussel-watch programmes could
be greatly expanded.
4. BIOLOGICAL STUDIES AT POPULATION LEVEL AND ABOVE
It can be argued that although effects may be shown at the level of the
individual or below, in dealing with organisms other than Man it is only when such
effects have consequences at the population and community levels that the
pollution has any real significance (Mclntyre and Pearce, 1980). Yet it is
difficult to measure effects at these higher levels, and in particular to detect
a statistically significant change in population size which could be related
unequivocally to pollution. Indeed, there are almost no life-table data available
on naturally occurring non- commercial marine species so that little is known on
changes in survivorship (l x ) or mortality (EL) in response to pollutants. Linked
to studies of, for example, scope for growth, field studies of life -table data
could be a promising area of research.
Most field studies of pollution effects at the population level have been
directed to the plankton or the benthos. Fish may seem to offer a promising
approach, in view of the large amount of data on commercial species. However the
major impact on the stocks is undoubtedly due to fisheries exploitation, and this
makes it difficult to identify stock changes that can be attributed to pollution.
Thus, although a great deal of information is available from commercial fishery
research and statistics, it cannot be said that fish are particularly attractive
for pollution assessment studies at the population level, although they are much
used for chemical monitoring.
4. 1 The use of plankton
Plankton, on the other hand, is a primary recipient and target for the great
majority of polluting inputs. Reference has been made to the use of zooplankton
(oyster and echinoderm larvae) in bioassay experiments, and there are descriptions
of the employment of phytoplankton in rotating dialysis bags suspended from buoys
in situ for marine pollution monitoring (Jensen, 1980).
There are also studies of plankton in mesocosms - large enclosures which
can be treated with a variety of contaminants and the populations studied over
periods of at least several months, thus introducing ecological realism to the
experiment. Mesocosms can range in size from a few cubic meters to as much as
1,300 m 3 and the design can be varied to suit a diversity of facilities,
requirements and costs. This approach may be recommended as a useful adjunct to
other studies. It can help in the interpretation of field observations and in
suggesting new ways of looking at problems. The use of mesocosms is reviewed by
Grice and Reeve (1982) and Kuiper and Gamble (1988).
However, direct field studies of effects on plankton are a different matter.
Even although the initial release of contaminants may be in the pelagic zone,
dilution and dispersion tend to be rapid there, and organisms will be carried away
from fixed sources of input. One consequence of this dilution and transport is
that impact on plankton is likely to be substantially less than on sessile benthic
organisms .
This does not apply in areas where water exchange with the open sea is
restricted and flushing is poor. In such conditions effects on plankton may be
expected and indeed are well documented, particularly in relation to excessive
input of nutrients (Relevante et al. . 1985). This results initially in enhanced
plant growth, culminating in unusual plankton blooms (sometimes toxic), and in a
change in the structure of the phytoplankton community, with smaller flagellates
replacing larger diatoms as the main components. Changes in the zooplankton may
also occur. Benovic et al. (1987), working in the northern Adriatic, recorded
changes in the hydromedusan faunas, with significant reduction in those components
(anthomedusae and leptomedusae) which have bottom-dwelling (hydroid) phases in
their life cycle. This is correlated with reduced oxygen in bottom waters
associated with eutrophication. Thus, changes in the structure of plankton
communities in certain areas can be good indicators of eutrophication.
But, in general, impacts of pollution in the water column are best studied
in relation to the nature of the input. First, in the context of point sources
of input, and second when the input is diffuse. Point source input is
particularly relevant to plankton in the case of ships dumping at sea, and
releasing large amounts of material, usually industrial wastes or sewage sludge,
into surface waters. For example, studies at an ocean dump site tor
pharmaceutical wastes off Puerto Rico showed an immediate change in the
phytoplankton community structure at the site, but no persistent long-term changes
were observed (Murphy al. . 1983). Further, looking at zooplankton, a survey
of industrial wastes discharged in deep water off the east coast of the United
States (Capuzzo and Lancaster, 1985) showed that only a small percentage of the
community was affected, and that long-term consequences to the zooplankton
populations were negligible. Another situation relevant to plankton is an oil
spill producing a major surface slick. There are a number of studies in which the
plankton under such slicks have been examined. Under a fresh slick from a large
spill in the most extreme conditions, organisms and pelagic eggs may be coated
with oil and there will be toxicity effects. However recovery can be complete in
a matter of days (Davenport, 1982). If there are no significant effects from
these discharges, it is most unlikely that any impact from diffuse inputs could
be detected. For these reasons plankton is not considered further in the manual.
5. THE USE OF BENTHOS
As suggested by the discussion above the most suitable programme for
examining the effects of pollutants on marine systems will be an analysis of
effects on benthic assemblages. Such assemblages are widely used because:
a. the organisms are largely sessile and must therefore tolerate the
pollution or die.
b. the assemblage integrates effects of pollutants over time.
c. a wide range of taxonomic diversity exists with upwards of 100 species
per sample for both macrofauna and meiofauna.
d. there are many examples of such assemblages showing effects of
pollutants [e.g. oil pollution (Davies et al. . 1984; Reiersen et al. .
1989; Gray et al. . 1990); organic enrichment (Pearson and Rosenberg,
1978; Mirza and Gray, 1981; Bellan, 1985; Warwick e &L_, 1987; Bellan
and Bourcier 1988, 1990); tannery effluent (Zenetos and Papathanassiou,
1989).
5.1 Choice of sample site
In the case of point-source pollution sampling will be done in the vicinity
of the discharge and although the details of the sampling programme will require
attention, the selection of the site in general will be obvious. However, when
studying diffuse sources, careful site selection is essential, since a major
problem in analyzing the effect of pollutants at the population level and above
is the difficulty of separating 'nuisance' environmental variables from pollutant
effects. Depth variability and grain size variation are typical
1 nuisance' variables . Where possible such variables should be held constant by
sampling at constant depth or within a narrow range of grain size. If this is not
possible it is important to match, for their physical variables, control sites
with impacted sites so that valid statistical comparisons are possible. A sample
site for the study of non-point source contamination is often difficult to define.
The objective should be to demonstrate whether a delimited geographical location
is more impacted than a control site. A site may, for example, be defined as a
lOOxlOOm area from within which adequate replicates can be taken.
5.2 Acquisition of data and preliminary investigations
The design of programmes for quantitative and representative sampling of
benthos is always a difficult task, and the more information there is on a
selected area before planning, the better the sampling programmes that can be
designed. The following information is needed for proper planning of quantitative
benthic studies.
(i) Ba thyme trie and geomorphological data for the investigated area
which may be available from existing documents, compiled into a
basic chart of the area. If such a compilation is not adequate for
the presentation of major geomorphological formations of the
submersal coastal slopes and of the plain sea bottom, additional
echo -soundings along critical transects in deeper waters and
orientative mapping by divers for coastal hard bottoms should be
done.
(ii) Sedimentological data from all available sources, including
navigation charts, plotted in the form of convenient histograms on
the basic map. By simple interpolations a map of the topographical
distribution of the major Sedimentological types within the
investigated area can be made. This information, combined with the
knowledge on the distribution of distinct types of water masses
within the investigated area constitutes one of the most important
elements in planning quantitative sampling programmes. Therefore,
it is advisable to complete Sedimentological studies of the
investigated area (if such data are not already available) before
the final setting-up of benthic sampling programmes, although both
samplings are usually carried out at the same time for practical and
economic reasons.
(iii) Oceanographic data are essential on distribution of water masses and
their movements as well as the trophic conditions in the pelagic
environment of investigated areas.
(iv) Any known inputs of pollutants should be examined with a view to
charting their spatial distribution within the investigated area.
This has to be done in order to select a suitable area for study and
to identify its extension and limits. In some cases information is
provided by the data on, for example, spatial distribution of
coliforms and detergents, which can serve as tracers for the marine
distribution of sewage and for the majority of mixed industrial
effluents. For industrial effluents which do not contain these
tracers, the detection of effluent distributions will require
analyses of specific communities. The most useful information
generally and for the latter cases in particular, is provided by a
knowledge of prevailing currents and other movements of water masses
within the investigated area, indicating most probable distribution
of pollutants.
(v) Qualitative data on types of benthic communities and their biota,
alone or together with the above information, form the basis for the
design of sampling programmes. Therefore all existing information
should be compiled and brought up-to-date by preliminary benthic
investigations, carried out by qualitative dredging on soft bottoms
and by direct diving observations and collecting on hard bottoms.
Observations and underwater photographs made by divers are also
extremely useful. The divers' information can be supplemented by
the use of remote cameras or underwater television equipment, and
indeed for deeper bottoms (over 50 m) , where divers cannot work,
these will be the only photographic opportunities.
5 . 3 Design of a benthic sampling programme
The ultimate objective of the programme will be to detect any change in the
benthos, spatial or temporal, in addition to that due to natural variability, and
to attribute the change to its cause.
If there is a point source such as an effluent pipe, a dumping vessel or
an oil platform, and therefore a relatively small area of impact with a strong
gradient, then the preliminary survey will suggest roughly the direction and
extent of any dispersion, and will indicate the most appropriate layout of
sampling stations. This may be in the form of a grid, covering the area or in
lines arranged to sample the range of the gradient.
On the other hand, if diffuse inputs are being studied and therefore an
JLU
extensive area is involved, then as a result of the preparatory work, an initial
survey should be carried out to map the extent of the various types of habitat
within the area. It is obvious that sampling methods will vary with habitat, the
optimum approach for sea-grass beds will be different from that for soft sediments
and different again for rocky situations.
Within each relatively homogeneous habitat it is recommended that a
stratified random sampling programme be used. In such a programme the expectation
is that the fauna or flora responds to key environmental variables. For example
in a sea- grass bed, depth may be an important variable so that sampling should be
stratified for depth. In the case of soft sediments it is known that grain size
has a strong influence on benthic macro- and meiofauna so the stratification
should be for grain size variations. The area is mapped and the various grain
size distributions plotted in.
An example follows, taken from Elliott (1971), showing how to plan a
stratified random sampling programme for a benthic survey. It is planned to
sample an area of 200 m 2 with a grab taking an area of 0.05 m. Potentially
therefore, there are 200/0.05 - 4,000 sampling units within the area. A
preliminary survey shows that the bottom is very heterogeneous. Since it is known
that grain size variations could be important in determining species
distributions, the sediment is mapped. Sampling should ultimately be done with
equal intensity on each type of bottom. This is called proportional allocation
of samples. Here an even coverage of 10% is given to each area i.e. 40 samples
total, a not unreasonable number.
The preliminary survey shows that gravel (nl) covers 1000 sampling units,
coarse sand (n2) 500, sand (n3) 1500, fine sand (n4) 800 and mud (n5) 200,
totalling 4,000 sampling units.
The 40 samples are then allocated in proportion:
nl - 1000 * 40/4000 - 10 samples
n2 - 500 * 40/4000 - 5 samples
n3 - 1500 * 40/4000 - 15 samples
n4 - 800 * 40/4000 - 8 samples
n5 - 200 * 40/4000 2 samples
As to the placement of samples, ideally the whole area is divided up and
each potential sampling unit is given a number, picked out from a table of random
numbers .
Another method of allocating samples within a stratified random approach
is optimal allocation, which provides for more samples to be taken where there is
high variability. An example of such an approach is shown in Appendix 1.
5.3.1 Size and number of replicate samples per site
The general rule for sampling is that many small samples are better than
few large ones. The reasons for this are that with many small samples a greater
coverage of the sample site is achieved, a better estimate of the spatial
dispersion of species is obtained and there are a greater number of degrees of
freedom for statistical analyses. However, the size of sample should not be
reduced to such an extent that so-called 'edge effects' (disturbance of the sample
by the edge of the sampler) dominate. More often than not the size of sample is
prescribed by the type and size of the gear available. It should not be assumed
however, that a particular grab (or plankton net) is the appropriate size for a
fiven population simply because it is available. Most grabs were developed as
ractions of 1m 2 and are not necessarily appropriate for all situations. For a
general guide to marine examples for determining sample size the reader is
referred to Elliott (1971) and Venrick (1978).
For statistical analyses an adequate number of replicate samples must be
taken. In practice this means a minimum of 2-3 replicates and with benthic
sediment samples 5-10 replicate samples is common.
There are a number of more or less objective criteria for determining the
number of replicates required. In sampling sediments an estimate of the total
11
number of species within a given area may be needed. A species-area curve is then
plotted of the cumulative number of species against number of samples (Figure 1).
From the shape of the curve an estimate can be made of the number of replicates
necessary to obtain an acceptable % of the total number of species. In this
example for C 4 five samples will give approximately 70% of the total number of
species.
AREA SPECIES CURVES, COMPARATIVELY LAKE OF TUNIS'
NORTH ADRIATIC (A) AND WITHIN GULF OF TRIESTE (B)
Figure 1. Various types of area/species curves from "normal" (curves C A , Cc, S x )
and pollution or estuarine "stress communities" (curves C l -C 3t K lt JC 2 and
S 2 ) . (From Stirn et al. . 1975, with kind permission of Pergamon Press
Ltd. , Oxford)
It may also be useful to know the number of replicate samples needed where
the primary aim is a study of the population dynamics of one single species. The
interest is in estimates of the population mean number and its variance. A simple
method to determine the number of samples required is to take five samples and
calculate the mean and variance. Take five more and calculate the mean and
variance for all ten samples and repeat until the mean and variance are stable.
The minimum number of samples which give a stable mean and variance should be
used.
A more elaborate method is to decide on an acceptable error of the estimate
of the population mean and use this to calculate number of replicates. An example
is shown in Appendix 2 .
In cases where the interest is in obtaining population estimates for a given
12
species the number of samples to be taken is often large. For example in a study
of the bivalve Mva arenaria in the Oslofjord, using the above formula, 30-35
samples were taken on each sampling occasion and the required number of samples
was calculated in the field (Winther and Gray, 1985).
In conclusion, the number of samples to be taken depends greatly on the
question asked. If the interest is related to the number of species, or species
diversity, then samples must be taken to obtain the maximal number of species.
For studies on the dynamics of one or a few species the formula above may be used.
5.3.2 Temporal sampling frequencies
The frequency of sampling depends on the question being asked and on the
amount of information already available. If there is no background knowledge of
the area under study then samples at regular intervals over a year are needed to
ascertain the seasonal changes in the assemblage. In practice this usually means
sampling monthly. Knowledge of seasonal changes can reduce the sampling effort.
In pollution studies it is often important to know how the assemblage changes over
time in response to the pollution load. It may be a waste of effort to sample
seasonally and measure recruitment of juveniles of species which will die later
in the season. If the interest is in changes from year to year then sampling at
seasons where the lowest abundances occur (often winter) would be optimal. Figure
2 shows data for a subtidal hard bottom species sampled seasonally for a number
of years. Clearly for this species the same trends will be in evidence whether
sampling is at times of minimal abundance (winter), maximal abundance (summer),
or takes the mean of all seasons. So a rationalized programme would sample once
per year in mid-winter if the interest is in year-to-year fluctuations.
Ciena intestinafe
(data from Lundaiv)
1000
100
E
10
1970 1971 1972 1973
Figure 2. Ciona intestinalis populations in Gullmarf jord, Sweden studied using
stereophotographic methods. Jagged solid line: actual data measured
approximately monthly. Smooth solid line: mean abundance. Broken line:
minimal abundance. Broken and dotted line: maximal abundance. (Data
from Lunddlv, unpubl.)
5.4 Sampling and processing methods
Decisions on the methodology and equipment will depend on the aims of each
specific exercise, on the nature of the habitat involved and on the staff and
facilities available. Each individual scientist tends to have his or her own
preference for equipment and procedures, and any given laboratory may have its own
traditional approach, determined partly by its research history, As a result
there is a great diversity of methodology and for any single study it is usually
13
appropriate that the investigator should select the approach with which he is best
equipped and most comfortable. However, in the context of collaborative or
regional sampling programmes or international surveys, the use of standard, agreed
methodology is important if results from different laboratories are to be linked
and compared. This is discussed later in more detail.
5.4.1 Intertidal sampling
A few parts of the Mediterranean Sea have regular tides (eg the Gulf of
Gabes, upper Adriatic) with an average vertical amplitude of 80 cm, but for the
most part tidal oscillations are extremely small, with vertical amplitudes of less
than 40 cm, resulting in narrow intertidal areas. Indeed, in the Mediterranean
vertical divisions in the marine environment are usually described in terms of
"zones". The region above the level not constantly covered by the sea is referred
to as the mid- littoral zone (divided into an upper sub -zone which is wetted only
by waves and a lower sub-zone, covered at high tide and wetted by waves only when
the tide is low) and the suppralittoral zone which is wetted only by spray and
where immersion is exceptional (Augier, 1982). The species composition of these
communities varies considerably in different parts of the Mediterranean. The
major components are calcareous and soft red algae, some brown and green algae,
along with Intertidal species of, for example, molluscs and cirripeds. Where soft
shores occur the deposit is largely of sand, and the macrofauna (mainly some
burrowing polychaetes and amphipods) is low in both species diversity and
population density. The intertidal zone, narrow though it is, is extensively used
for recreation and tourism and is highly sensitive to contamination from the land,
so it will be important to study it in the context of pollution.
On rocky shores, the great diversity of habitat, including pools, exposed
rock, sheltered crevices and the undersides of stones, makes it difficult to
collect samples which are adequately representative of wider areas, and separate
estimates for each distinct habitat will be required. A square frame of wire can
be useful in defining the study or sampling area, and organisms either counted or
collected, and the percentage of the area covered by organisms estimated. Frames
of 1 ra 2 or 0.25 m 2 are frequently used, but on very irregular surfaces or for small
organisms, frames of 316 x 316 mm (0.1 m 2 ) are more suitable. For small species
such as barnacles which may occur close together in large numbers, even smaller
areas are appropriate, and a piece of thick perspex 100 x 100 mm (0.01 m 2 ) , etched
with a grid of 10 mm squares facilitates counting. Counting or sampling is
usually done along traverses with stations at regular intervals or at specific
tidal heights.
On soft shores representative sampling is easier. Undisturbed cores of
sediments can be collected by pushing tubes of plastic or metal into the sand, the
diameter of the tube being chosen depending on the volume or depth of sample
required. For larger samples a square frame of sheet metal, usually enclosing an
area of 0.1 m 2 or 0.25 m 2 is driven into the sand and the deposit dug out to the
required depth. The sand is then sieved through a mesh of 0.5 or 1.0 mm depending
on the size of the particles and the nature of the results required. A detailed
discussion of intertidal methods for both hard and soft shores is given in Price
et al. (1980).
5.4.2 Subtidal sampling: Ships and shipboard equipment
For shallow water work in the infralittoral zone close to the coast small
boats of 7-10 m length, with 20 to 30 hp engines are usually suitable and can
operate light gear-dredges, beam trawls and even small corers and grabs, which can
be hauled manually using ropes. Positions can be determined from marks on the
land.
For most subtidal work, however, larger ships are required that are fitted
with winches suitable for hauling wire ropes for dredging and wire for grabs and
corers which are operated vertically. For trawling and dredging, warps of from
12 to 24 mm diameter are used, and on the shelf a length of warp about 2*1-3 times
the depth of water is usually used. When operating grabs of 100-150 kg,
galvanized steel wire of 6-8 mm diameter is appropriate.
If a vessel built for research is available the required equipment will be
on hand, but if it is necessary to charter another type of vessel, it is important
to ensure that cranes or booms and winches with appropriate wire are available,
that relevant navigational facilities are fitted and that there is a suitable
echosounder. A convenient source of running seawater on deck is required and
there should be sufficient free deck space for handling samples as well as bench
space under cover for processing, and space for storage.
5.4.3 Subtidal sampling on hard bottoms
Rocky bottoms present the most difficult problems for remote sampling. If
not too steep or uneven they can sometimes be surveyed using a heavy duty dredge
which may provide at least qualitative samples. For example, a naturalist's or
rectangular dredge with a 12 mm nylon bag is suitable. Due to rocky or encrusting
irregularities or hard bottoms, the cables and other gear connected to the dredge
must be strong enough to hold forces up to 1000 kg. The dredge is equipped with
a weak link (Figure 3) consisting of several turns of twine (for heavy duty, three
turns of 8 mm manila rope) which breaks if the dredge is anchored or stuck between
rocks, allowing the arms to open and free it.
Figure 3. Naturalist's dredge. Note the position of a "weak link". (From Holme
and Mclntyre, 1984, with kind permission of the International
Biological Programme, London)
However, in general, remote sampling gear such as dredges, grabs or corers
are not appropriate for hard ground. Underwater photographic or television
cameras lowered from a ship can provide useful information, but the most suitable
approach is the use of diving. This is dealt with in detail in Section 5.4.5.2.
5.4.4 Subtidal sampling on soft bottoms
5.4.4.1 Qualitative and semi -quantitative sampling
For subtidal sampling on soft bottoms a wide variety of equipment is
available, allowing for a diversity of requirements. Towed gear such as dredges
and trawls provide qualitative and sometimes even semi -quantitative material.
Beam trawls are available in several different forms (Figure 4) but basically they
all consist of a long net, the mouth of which is held open by a rigid beam with
metal runners at each end. The lower leading edge of the net is usually weighted
or attached to a chain which curves back behind the upper leading edge of the net
attached to the beam, thus preventing the escape upwards of mobile organisms
disturbed by the ground rope. The Agassiz trawl is essentially a symmetrical beam
trawl without a leading edge and with the net attached to two metal shoes so that
it can be used either side up. It is thus suitable for deep water sampling when
the landing of the gear on the bottom is difficult to control.
15
Figure 4. Beam trawls for qualitative sampling, (a) towed stow net; (b) 'Keitel'
of Curishe and Frische Haff (both are lagoons on the southern coast of
the baltic) ; (c) Japanese beamtrawl; (d) modern European beamtrawl for
shrimps. (From von Brandt, 1981, with kind permission of the author)
Dredges, which are simply heavy metal frames fitted with a bag or coarse
net, have already been discussed for hard bottom sampling and are also suitable
for soft sediments. They are easily made in a wide range of sizes and weights,
and the collecting bag can also be adapted in design and construction to meet
specific sampling needs. Light dredges can be hand-hauled from small boats, while
for deeper water, heavier gear is required for operation by winches. On silt-clay
grounds on the shelf or beyond, the anchor dredge of Sanders can provide large
samples. As shown in Figure 5, it has two angled digging edges with a heavy
horizontal plate between them which determines the digging depth. Another semi-
quantitative instrument widely used in the Mediterranean is the Charcot dredge
(Picard, 1965).
The otter trawl, as used by commercial fishermen, has the net spread open
by two otter boards. The trawl is shot on twin warps either over the side or from
the stern and is reeled in using a double-barrelled winch. A variety of trawl
gears is now available and a considerable range of ancillary equipment makes
trawling, although still not fully quantitative, a highly sophisticated operation
and one which, particularly at shooting or hauling, should be done by experienced
specialists.
While trawls and dredges can be made at least semi-quantitative by
standardizing as much as possible the condition and duration of towing, they are
basically qualitative sampling devices. However, being mostly relatively simple
gears, they can often be used in circumstances when more sophisticated equipment
is not appropriate, and are invaluable in providing an initial indication of the
general nature of a habitat and its fauna and flora.
5.4.4.2 Quantitative sampling
Reasonably quantitative sampling on soft bottoms is possible, and grabs and
corers are appropriate, - instruments which are lowered vertically on a warp, the
grab having jaws which 'bite' out a volume of sediment, while corers penetrate the
deposit and on being hauled, carry a plug of sediment with them.
(a) Macrobenthic infauna. Many methods and types of sampling gear are available
and a comprehensive review is given by Eleftheriou and Holme (1984) who list and
discuss more than 20 samplers. Some of these are highly sophisticated and require
specialist back-up equipment or the assistance of divers and can be most effective
if such support is available. The ultimate choice of sampler will depend on the
detailed requirements of the exercise as well as on the working conditions and the
nature of the sediment. It is advisable to use the simplest gear that will
provide a satisfactory sample and for general collecting and survey work one of
It)
caufamc i
NET
Figure 5. Anchor dredge Sanders type (From Sanders et al. . 1965, with kind
permission of Pergamon Press, Oxford)
the several developments of the original Petersen grab - a weighted hinged bucket,
is recommended. The Van Veen model is virtually a Petersen grab with arms. The
Smith-Mclntyre added a frame, springs to drive the jaws into the sediment, and
trigger plates at opposite corners which ensure that the grab would not activate
until it was sitting squarely on the bottom. The Day grab is a simpler version
of the latter and is now widely used. These grabs usually cover a surface area
of 0.1 m 2 (although smaller and larger versions are available) and weigh around
30 kg empty. On soft mud a light grab will penetrate well but on hard-packed sand
it is important that the gear should be heavy enough to bite deeply into the
sediment rather than simply scrape the surface on lift-off. The digging depth of
the instrument should be at least 10 cm. In this context the volume of sediment
in a grab should always be measured, either by a graduated stick or by
transferring it to a marked bucket. Less than about 4 litres of sediment in a 0.1
2 grab indicates poor penetration, and if the volume of sediment is too low the
sample should be rejected. The handling of the winch which operates the grab is
also important. The wire should be kept as vertical as possible to ensure that
the instrument is set down and lifted up at right-angles to the bottom, and this
means that working in bad weather is difficult or impossible. It is often useful
to stop the winch for a moment just before the grab hits the bottom so that the
setting down is done as gently as possible to lessen the shock wave and reduce the
washing away of sediment. Hauling should be commenced as soon as the grab is
properly settled on the bottom, since any delay will increase the wire angle if
the ship is drifting, and the instrument will be hauled out obliquely giving a
reduced sample. It is important to haul very slowly until the sampler has left
the bottom.
In contrast to the principle of the grab, a corer is an instrument in the
form of a tube which penetrates the bottom usually by its own weight, and retains
a plug of sediment. Corers are usually smaller than grabs and are discussed below
17
in the context of meiobenthos, but one type of corer appropriate to macrofauna is
the box sampler, which is considered here. This was first described by Reineck
(1958) and consists of a rectangular corer supported in a metal frame. The corer
penetrates the bottom assisted as necessary by added weights, and a hinged cutting
arm swings down when the warp is hauled, closing the bottom of the tube and
retaining the sample. The instrument samples an area of 20 x 30 cm to a maximum
depth of 45 cm and weighs 750 kg in use. There have been several modifications
of this design, such as the spade corer of Hessler and Jumars (1974) which covers
an area of 0.25 m 2 . These instruments are most satisfactory in that they provide
relatively undisturbed samples to considerable depths, and the boxes containing
the cores can be removed for convenient study in the laboratory. Unfortunately,
box corers are not only expensive but also extremely heavy and very large, so that
they are difficult to work and require the facilities of a large ship.
(b) Nelobenthos. Because of their small size and high density, organisms of the
meiobenthos are best collected in small samples of sediment. This can be done by
subsampling from a grab haul but such a procedure is unsatisfactory for several
reasons, particularly because the down-wash of the grab will have disturbed the
surface sediments where the meiobenthos is often richest, and because the closing
of the grab and its passage to the surface will further disturb the contents so
that a representative sample will be difficult to obtain. Subsampling from a grab
should therefore be a last resort. A box sampler can provide a better possibility
for subsampling, but even here some of the objections listed for grabs are likely
to apply.
The most satisfactory approach is to collect a small sample dedicated to
meiofauna, and the ideal instrument for this is a corer. The best samples are
taken by divers operating a core tube manually but as indicated in section 6.4.5
there are significant constraints on diving work. The simplest remote corer is
still an open barrel instrument such as that designed by Moore and Neill (1930).
This collects good samples from muddy bottoms, especially if it penetrates a layer
of clay which acts as a plug to prevent the mud core sliding out. Often, however,
and always on sand, some type of core retainer must be fitted to close the lower
end of the tube. A more sophisticated instrument is described by Craib (1965).
This is a tube of 5-7 cm diameter which is mounted on a frame. When the frame
comes to rest on the bottom the tube is forced slowly into the sediment by
weights, controlled by a hydraulic damper, ensuring minimum disturbance of the
light surface layer of the deposit. Samples of 15 cm in length are obtained and
a closing device ensures that even samples from hard-packed sand are retained.
The apparatus, weighing 44 kg, can be handled from a small boat. A larger
version, employing multiple core tubes for use in deep water has been developed
by P.R.O. Barnett of the Scottish Marine Biological Association and is described
in Holme and Mclntyre (1984). The advantage of cores for meiofauna is that the
whole sample can be examined (split into layers if desired to study vertical
distribution of the fauna) without resort to subsampling which can cause errors.
For this reason the diameter and length of the core should be carefully selected
so that the effort in counting and processing the fauna is not too great. It has
been found that cores of 2-4 cm diameter are satisfactory for most purposes. In
areas of very dense population smaller diameters may be appropriate but if the
tube is too narrow, difficulties will be encountered in adequately collecting the
surface fauna (Mclntyre, 1971).
5.4.5 Other techniques
In several of the paragraphs above, reference has been made to special
techniques such as diving and underwater photography, which are at least useful
additions to a programme and in some cases the best or even the only adequate
approach. These are discussed in more detail below.
5.4.5.1 Diving
In shallow water, diving using self-contained underwater breathing apparatus
(SCUBA) allows detailed direct observations to be made, notes and photographs to
be taken, and in situ experiments to be set up and operated. It also permits the
collection of specimens and samples by hand, thus giving access to material from
habitats, such as rocky bottoms, where the use of remote sampling gear is
difficult or impossible. A comprehensive review is given by Gamble (1984).
18
In view of the value of this approach, it is important to recognize the
limitations. Diving can be extremely dangerous and it is essential that those
involved should be adequately trained, be fully aware of the safety requirements
and should observe the rules, as set out in the manuals and codes available (eg
NOAA, 1979 and Underwater Association for Scientific Research Ltd. 1979). When
working in sewage -polluted waters, divers should be vaccinated appropriately for
protection against pathogens likely to be encountered. Divers should operate as
part of an experienced team and should never work alone, indeed in some countries
this is illegal. The main restriction to SCUBA diving is depth. The time of a
dive on air to deeper than 10 m is limited by decompression considerations and
nitrogen narcosis restricts air diving to about 50 m. Temperature also limits the
time a diver can spend underwater and visibility can be another constraint, since
suspended particles, either biogenic or inorganic, can reduce transparency and
light to virtually zero.
Divers can work free in the water, or can operate from towed gear or driven
vehicles such as submersibles . The simplest gear is the underwater sledge which
can be towed, for example, along with a trawl.
Whenever possible the diver should log his observations in situ to prevent
inaccuracies which might arise from relying on memory for later recording. A
camera is obviously of great assistance and its use is discussed below. More
immediate recording can be done underwater by making notes on plastic board using
a graphite pencil. For convenience the pencil can be attached to the board by a
cord and the board should be fitted with a wrist strap. The use of speech is also
possible and divers can either talk into a cassette tape recorder in a waterproof
housing strapped to the aqualung cylinder, or communicate directly with the
surface vessel. In this context bone conduction microphones held tightly against
the diver's skull by the suit hood are appropriate (Main and Sangster, 1978).
However, modification of the usual diver's mouthpiece is required if
understandable speech is to be produced, and some practice is required in
interpretation. In skilled diving teams the hand signal speech of dumb persons can
provide a useful means of communication.
In spite of the limitations already discussed the diver, in collecting
samples, has certain major advantages over remote gear used from ships. In very
turbid areas over fine sediments it may be necessary to operate by touch, but in
general the diver can examine the substratum and determine the exact location of
his sample; he can see the diversity of the bottom and will know how
representative his sample is; he can observe the reactions of organisms and be
aware of the escape of mobile species. The most straightforward sampling
operation is simply the picking up of flora and fauna by hand, using hammer and
forceps when necessary. Larger specimens can be placed in prelabelled mesh sacs,
polythene bags or jars, while smaller or delicate organisms can be collected by
suction devices. For some cryptic fauna, such as that of macrophyte holdfasts,
one approach is to collect the entire habitat, while encrusting or attached
organisms can be scraped off into containers. Mobile species can often be induced
to leave burrows or crevices by squirting in dilute formalin or bleach.
Apart from the direct picking up of organisms by hand, the simplest
collecting by divers is done with a corer on soft bottoms, using a small tube
(5-10 cm diameter is popular) of plastic or metal which is pushed or hammered into
the sediment, and sealed with a rubber bung before transportation to the surface.
Care must be taken (Mclntyre, 1971) to ensure that the flocculent material lying
on the sediment surface is collected, since this often contains a significant
proportion of the smaller fauna, and is easily lost if small diameter cores are
used. Another useful piece of hand-held equipment is the suction sampler, which
can readily be devised to pick up small or delicate organisms (Tanner et al. .
1977) .
Divers can also make a major contribution by working along with surface
vessels to position and operate gear lowered from ships. A range of hydraulic and
air-lift samplers are available which are particularly suitable for such
collaborative work (Elephtheriou and Holme, 1984).
Two further activities in which divers can play important parts are survey
operations and underwater experiments. When studying pollution from a point
source, surveys along transects can be particularly relevant and in areas of hard
19
bottom divers may offer the best or even the only possibility of obtaining data.
The transect can be defined by setting out a non-buoyant line on the bottom. If
the seabed is steep the line should be fixed to the bottom, and it can serve
additional functions (eg for storing and recovering samples) if buoys and clips
are attached along the line (Bailey et al. . 1967). Quantitative data can be
obtained by swimming along a transect and counting individual organisms within an
area defined by the length of a horizontal rod (usually 1 m) held at right angles
to the transit line. A more accurate method involves the use of quadrats, within
which the organism can be visually assessed or photographed, or all the material
in the quadrat can be collected. The quadrat is usually defined by laying a rigid
frame on the bottom, and frames of up to 1 m 2 can be used without inconvenience,
but the size of the frame required will depend, among other things, on the type
of habitat under study. Larkum et al. (1967) required 1 m 2 to differentiate
significantly between plant assemblages at 30 and 45 m off Malta.
Finally, divers can make a unique contribution by their ability to set up
and service underwater experiments, which can be highly useful in pollution work.
Experiments can range from the study of in situ respiration of organisms or
communities (Boynton et al. . 1981; Loeb, 1981) to the setting up of underwater
enclosures for the manipulation of populations and habitats.
In summary, divers can play a major role in ecological studies, but because
of the short time they have available underwater, it is important that they should
optimize their activity by careful advance planning. In addition, the dangers of
their operations must be fully recognized and allowed for.
5.4.5.2 Photography
Photography and television are now widely used in benthic studies and offer
the considerable attraction of non- destructive sampling (Holme, 1984). In the
intertidal zone photographs, particularly in colour, provide valuable records of
the distribution of plants and animals, while aerial photography allows a more
extensive coverage and assists in mapping and defining shorelines and submerged
reefs and banks.
It is however underwater that photography, video and TV have a special role
to play, used either directly by divers or remotely operated. They may be
employed in conjunction with other sampling approaches or used as the major aspect
of an investigation. For estimating epifauna on hard bottoms or for enumerating
large, sparsely distributed organisms these techniques are invaluable, and
sometimes provide the only possible means of studying animal behaviour in situ.
Also, a photographic or TV survey of an unknown area can be of great assistance
in planning a detailed programme.
One very successful technique is to use stereo-photograrametry (Lunddlv,
1971). Here a pair of cameras are mounted on a frame and set to take pictures at
fixed distances from the substrate. The stereo pairs of pictures so obtained can
then be analyzed in a stereo- comparator and 3-d images reconstructed. This
technique is especially useful for subtidal studies where growth rates can be
measured, in the third dimension, down to 0.5 mm. Figure 6 shows the frame. Used
underwater a bar is attached to the substratum with notches at fixed distances
along the bar. The camera frame is hung on the bar and replicate pictures can
then be taken at sites along the bar (Christie et al. . 1985) .
Once photographs have been obtained it is possible with modern techniques
to transfer the data directly to the computer. The photograph is placed on a
digitizing board and then using commercially available software the community can
be sampled, either under fixed points (the point sampling method to give areas of
coverage) or the areas of dominant species can be traced using the digitizer. The
data are stored directly in the computer for further analyses of spatial and time
series changes in abundance patterns of individuals and species.
Detailed studies of growth rates of individual species can be made, provided
the exact same area is sampled at each time interval. The photograph is placed
on the digitizing board and from the coordinates given, the computer stores the
data for each individual and can be programmed to plot, for example, growth rates
from these data. Examples of the use of such techniques can be found in We thy
(1984).
Figure 6. Stereophotographic frame used in conjunction with underwater camera;
c- holes for insertion of cameras; m- frame mounting point on bar
bolted to study site (from Lundalv, 1971)
Technology in this area is expanding rapidly. It is now possible to use
cameras at remote sites, which digitize pictures and relay these by radio directly
to storage units at the home base. The digitized pictures can be enhanced using
software developed for processing satellite images. Automatic recognition or
dominant species should be fairly straightforward so that it should soon be
possible to register automatically changing dominance patterns at remote
underwater sites (Christie, 1983).
A new and important development is the sediment profile imagine camera
developed by Rhoads and Germano (1982) for remote ecological monitoring or the sea
floor (REMOTS). The principle is that many sea floor processes can be
reconstructed from sedimentary and biological features found in the upper 20 cm
of the sea floor. The REMOTS camera allows high resolution imaging of these
features by means of in situ photography via an optical prism which sections the
sediment. The negatives obtained are analyzed rapidly by a computer image analysis
system and measurements of grain size, boundary roughness, thickness of dredged
material, depth of RPD layer, epifauna, , tube density surface aggregations of
bacteria and bioturbation can be measured. The system in its latest development
(Science Applications International Corporation, Admiral's Gate 221 Third St.,
Newport, R.I. 02840, U.S.A.) can be used down to 4000 m and can sample up to 100
sites per day.
5.4.6 Treatment of samples
Samples from dredges or trawls will usually consist of macrofauna
relativelyfree of fine sediment, but often associated with coarse material
including pebbles and rocks, which should be examined for encrusting organisms
before it is discarded. Large organisms, whether attached or free, should be
examined first in seawater. Some species (such as actinarians and other soft
cnidarians, turbellarians , opisthobranchs and other molluscs without exoskeletons,
nemertines, echiurids, priapulids, sipunculids, and enteropneusts) contract on
preservation and may radically change their body form. These require to be
anaesthetized (using menthol crystals or MgCl 2 up to 4% concentration) before
preservation (see Steedman, 1976 and Lincoln and Sheals 1979 for detailed
discussion of narcotizing agents). Most taxonomic groups, however, can be
preserved and stored in the same solution as used for primary fixation (e.g. 5-8%
formalin in seawater) . To prevent the formalin solution from becoming acidic and
dissolving calcareous structures, use a buffering substance such as hexamine
(about 8g hexamine per litre of 2% formalin solution). Sponges and halothurians
should be stored in 70-80% alcohol.
21
Grab or core samples, on the other hand, will be obtained usually in a large
volume of sediment and must be processed before preservation. Macrofauna and
meiofauna require different approaches.
5.4.6.1 Macrofauna
Initial separation of the organisms from the sediment in the field is
usually done by sieving through a screen, after which the residue on the screen
is transferred to sample jars, preserved and labelled. Further separation from
any remaining sediment ('sorting') prior to final identification and processing
is done in the laboratory.
Screens are made from high quality stainless steel or bronze gauze at the
bottom of stainless steel or plastic frames 15 to 25 cm high, depending on the
sieving procedure to be applied. The free surface of the screens should be about
1,000 cm 2 , or 30 x 30 cm; the outer surface must be reinforced, for instance, by
a stainless steel cross. If using a series of screens it is convenient to
construct frames in the form of drawers to be placed in a rack- like stand.
Ideally, they should be made to fit completely into a large plastic or enamel tray
so that all the screened material can be shaken down at once from the sieve into
the tray.
The total sample or portions are transferred from the grab into the upper
sieve and then the sieving is done by washing the material with gentle jets of
seawater, shaking by hand and separating agglomerations. Fixed sprinkler- tubes
or flexible heads, such as a shower nozzle, must be used for washing. For large
sampling programmes, and if working in heavy seas, more robust systems for sieving
operations are recommended, such as the Holme's hopper (Holme and Mclntyre, 1984)
shown in Figure 7. If no running seawater is available, the simplest sieving
method is to transfer a portion of a sample into a sieve or tightly connected
series of screens placed in a fairly large bucket with seawater, and to shake
continuously until the sediments are washed out.
It should be stressed, however, that all procedures described above may
damage delicate organisms, particularly polychaetes. Therefore, Sanders
extraction methods (Sanders et al. . 1965), are recommended for high-level sampling
programmes (Figure 8). The sample is washed by putting it in a large container
which has a spout near the top, much like a coffee pot. A large diameter water
hose (e.g. 4 cm) is pushed down into the sediment, and a large volume of water
running at a low velocity is pumped through the sediment. The resulting
suspension of animals and fine-grained sediment pours out the spout and then
through the mesh screen. The animals are retained on the screen. Large animals
are immediately picked out and preserved. At the end of the washing process there
are three fractions: animals taken out, the fauna retained by the screen and a
coarse fraction remaining in the container, consisting of coarser sediments
(heavier organisms such as molluscs). The three samples are preserved separately.
This method is time consuming but it is also extremely gentle, and in general the
animals are well preserved and relatively undamaged.
The mesh size of the screen used will be determined by the purpose of the
investigation and the type of sediment encountered. Usually a mesh of 1.0 mm -
0.5 mm will be required, but smaller meshes may be used with very fine sediments
or when juvenile stages of the fauna are required.
When sorting preserved samples, it should be remembered that formalin is
toxic and probably carcinogenic. It should therefore be handled with great care
and a means of waste air exhaustion should be provided for all laboratory
procedures. For sorting, the samples should be washed thoroughly with tap water
to ensure that sorters are not exposed to formalin vapour. Sorting may be
facilitated by staining the sample first, especially if many small animals are
present. Rose Bengal is a suitable stain and the following procedure is
recommended (ICES, 1990):
wash the sample free from the preservation fluid by using a sieve with
a mesh size smaller than 0.5 mm.
allow the sieve to stand in Rose Bengal stain (1 g/dm 3 of tap water +
5 g of phenol for adjustments to pH 4-5) for several minutes with the
sample well covered.
wash the sample until the tap water is no longer coloured.
If biomass determinations are required, this can be done using wet weight,
dry weight or ash- free dry weight, from either fresh or fixed material. For more
detailed work, energy content or equivalents of carbon, nitrogen or phosphorus may
be determined, but fresh material only should be used for these measurements.
Figure 7. Holme's hopper for sieving benthic samples. P - pipes supplying jets
along top of hopper; H - side-wall of hopper; R - retaining wall wall
at side of base (B) ; T - spout; G - rising gate; S - short legs
supporting hopper off base; L - legs; - sediment seen through gap
between hopper and base. (From Holme and Mclntyre, 1984 with kind
permission of Cambridge University Press, Cambridge)
Figure 8. Overflow elutriation system. (From Sanders et al. . 1965, with kind
permission of Pergamon Press Ltd. , Oxford)
23
Fresh wet weight is to be preferred to formalin wet weight, but if the
latter has to be used, weighing should not be done until at least three months
after fixation (Brey, 1986).
The wet weight is obtained by weighing after external fluid has been removed
on filter paper. The animals are left on filter paper until no more distinct wet
traces can be seen. Shelled animals are generally weighed with their shells, the
water should be drained off bivalves before weighing. When shell -free weights are
given, the shell weight should be included in the data list. Echinoids should be
punctured to drain the water before blotting on filter paper. As soon as the
non- tissue water has been removed, the organisms are weighed with the accuracy
required (for adult macrofauna weighing to 0.1 mg is often sufficient). In case
tube-building animals have to be weighed together with their tubes, appropriate
correction factors should be established. Dry weight should be estimated after
drying the fresh material at 60C, or by freeze -dry ing, until constant weight (at
least 12-24 hours, depending on the thickness of material). Dry weights obtained
by lyophilization (freeze drying) are slightly higher than those obtained by oven
drying. For Mvtilus. lyophilized tissues weighed 10.9% more than oven dried
tissues (Gaffney and Diehl, 1986).
The use of ash free dry weight is recommended in routine programmes, since
it is the most accurate biomass measure (Rumohr et al . . 1987; Duineveld et al A .
1987). However, it destroys specimens, and the consequences of this should be
carefully considered. Ash free dry weight should be estimated after measuring dry
weight. It is determined after incineration at 500 C in an oven until weight
constancy (ca. 6 hours, depending on sample and object size). The temperature of
the oven should be checked with a calibrated thermometer, because there may be
considerable temperature gradients (up to 50 C) in a muffle furnace. Caution is
advised not to pass a certain temperature (<550 C) since then a sudden loss of
weight may happen due to the formation of CaO out of the skeletal material of many
invertebrates (CaC0 3 ) . This can reduce the weight of the mineral fraction by 55%.
This decomposition occurs very abruptly and within a small temperature interval
(Winberg, 1971).
Before weighing, the samples must be kept in a desiccator while cooling down
to room temperature after drying, as well as after removal from the muffle
furnace .
To estimate biomass from length or size measurements conversion factors may
also be used (Rumohr et al. . 1987; Brey et al. . 1988).
The above account of weighing procedures is drawn from ICES (1990).
5.4.6.2 Meiofauna
Extraction, separation and sorting of meiofauna, particularly if needed for
reliable quantitative investigations, present a difficult task, and the most
time-consuming part cannot be done with the naked eye: all operations, except the
extraction, must be performed under a binocular dissecting microscope. The most
recent accounts are given in Holme and Mclntyre (1984) and in the comprehensive
Introduction to the Study of Meiofauna edited by Higgins and Thiel (1988).
The samples taken from a substrate of homogeneous fine sand or mud sediments
can be treated relatively easily by the Swedmark method, illustrated in Figure 9.
The sample, stirred to break up lumps, is placed in a large vessel and covered
with 1 to 2 cm of seawater. The mud surface is then pumped into suspension, using
a large silicone- coated pipette, and transferred to a nylon sieve or series or
sieves (250/i, 62/i) , the largest having a diameter slightly less than the normal
size of the petri dish used. Sieving is done by gently rocking the sieve in
seawater, either in another vessel or in the original one so that the filtrate is
returned to the original sample. When this is complete the sieve is placed in
seawater in a petri dish so that the fauna can be examined under a binocular
microscope before being transferred from the sieves, when they are likely to be
damaged .
For meiobenthic samples of coarser and sorted sand (true interstitial
meiofauna) the Boisseau elutriation method in closed- system is recommended
(Figure 10). This method is more sophisticated but not time-consuming. The
LARGE SI LI CONE-
COATED PIPETTE
NYLON MESH
150 y OR LESS
ACRYLIC GLASS
NYLON SIEVE
PETRI DISH
Figure 9. Swedmark method for extraction of meiofauna. (From Rulings and Gray,
1971, with kind permission of Smithsonian Institution Press,
Washington)
sample is placed in the separation funnel and an equal volume of 6% MgClo solution
is added (for anaesthetization of organisms which tend to attach to sand grains).
After about 10 minutes, a continuous stream of filtered seawater is introduced
through the tap on the separation funnel. After 15 minutes of elutriation, the
tap on the tube above the filter is opened and the water allowed to drain through
the sieve (50 to 70/i mesh). The sieve is inverted in a petri dish and the
meiofauna washed off with a jet of filtered seawater. A large part of the light
fauna will be collected on the sieve, but heavier organisms, such as molluscs,
ostracods and foraminifers , might remain in the sediment residue so this has to
be examined microscopically. This method can also be satisfactorily used for the
elutriation of preserved samples; in this case an open-system or a continuous
stream of seawater can be applied, only the incoming seawater must first pass a
filter in order not to contaminate the sample.
For the treatment of very heterogeneous samples, such as those obtained on
hard or marl-detritic bottoms, a convenient, although not entirely quantitative,
method is the seawater-ice technique described by Uhlig et al. (1973) (Figure 11).
The sample is placed at the lower end of a large plastic tube tightly covered by
120 to 150/imesh nylon gauze, which just dips into filtered seawater in a
collecting dish. The sample is covered by a layer of cotton wool, and the tube
is filled with the crushed seawater ice. As the ice melts, motile meiofauna move
through the gauze into the collecting dish due to salinity/ temperature gradients
and the streaming action of the water of different densities. If the samples to
25
Figure 10. Boisseau type apparatus for elutriation of meiofaunal samples, closed-
system arrangement. (From Holme and Mclntyre, 1984, with kind
permission of the International Biological Programme, London)
PT
Figure 11. Uhlig's method for the extraction of meiofauna I - insulation
material; NG nylon gauze; PI, P2 - petri- or culture dishes; PT -
plastic tube; S - sediments; SI - seawater ice; SW - sea water; TH -
tube holder. (From Uhlig e al. . 1973, with kind permission of
Biologische Anstalt Helgoland, Hamburg)
be treated contain a significant amount of mud, silt or clay, it is advisable to
wash them on a 50 to 70/i screen before this treatment. Obviously, only living
samples can be processed by this method.
A similar principle but using light can be used for the extraction of
phototactic vagile meiofauna inhabiting seaweed. Clumps of algae are placed in a
glass jar with seawater and exposed to a strong light source (sun or lamp) from
one side only. The creatures assemble along its illuminated surface and can be
picked up by a pipette. Subsequent washing of the algae clumps and examination
with a low silver microscope is recommended for quantitative results.
A colloidal silver polymer (Ludox-TM) can be used to separate meiofauna from
sediment and debris. The standard technique is described by De Jong and Bouwman
(1977) and a more rapid method involving centrifugation is outlined in Mclntyre
and Warwick (1984).
As mentioned, the final separation and sorting of meiofauna can be done only
under a stereoscopic microscope. The best type of sorting vessel is a medium-size
petri dish, with marked lines on its outerbottom for better orientation while
scanning the surface covered by the sample. For separation of organisms,
capillary pipettes, fine needles, loops and watchmaker's forceps are needed. In
order to make them more clearly visible, and to differentiate biota from detritus
and sediment particles, treated samples should be stained with Rose Bengal after
being preserved. For this purpose 10 ml of the stock solution (1 g stain
powder/100 ml ethanol) is added to 100 ml of sample plus preservative.
5.5 Analyses of results
Methods of analysis for data obtained from both hard and soft substrata are
similar and there are no special methods that need to be applied separately. With
the widespread availability of powerful personal computers and sophisticated
statistical analyses in packages it is assumed that such facilities are generally
available.
Many packages exist which are suitable for statistical analyses. For PCs,
STATGRAPHICS, SAS, SPSS and SYSTAT are highly sophisticated general statistical
analysis programmes and relatively easy to use. However, in analyzing
species/site matrices for possible pollution- induced effects the best results are
usually obtained using a combination of, first, multivariate analysis, and then
the statistical programmes testing specific hypotheses. As yet there are no
generally available multivariate analysis packages, but one is under development
by Clarke and Carr of the Plymouth Marine Laboratory, UK, for the IOC/UNEP/1MO
Group of Experts on Effects of Pollution (GEEP) and many of the analyses mentioned
below were performed using this programme on a PC with hard disk, Hercules card
and math co-processor.
Whereas most biologists have some knowledge of basic statistics and analysis
of univariate data (e.g. diversity indices), multivariate methods may be less
well-known. In multivariate analyses the complete data matrix of numbers (or/and
biomass) of individual species at all stations are analyzed as a single data set.
Powerful methods exist, now adapted to PCs, and can be generally recommended as
having been shown to be the 'best' method of unravelling the complex effects of
pollutants on marine assemblages. Most multivariate analyses, however, have been
done on the fauna of soft sediments and the examples given are from this area.
Attention is drawn to methods that are likely to be different.
The raw data obtained from surveys will usually be in the form of two
matrices of sites and species, one of abundances of the individual species over
sites and the other of biomasses of the individual species over sites. The
environmental data for sites should be recorded in a similar matrix.
It is recommended that the user of this manual becomes familiar with common
computer-based data file systems such as Lotus 123 or DBASE IV. The raw data
should be entered in the format of the above-mentioned programmes so that transfer
to statistical analysis packages later is straightforward.
The following analysis protocol has been developed over a number of years
by many investigators (eg. Field et &LU, 1982), but was formally described as the
27
result of a workshop conducted in the Oslofjord, Norway, under the auspices of
GEEP and is fully described in Gray <g al. (1988):
(i) Multivariate statistical analyses are used to discriminate between
sites based on their faunal (or floral) attributes using
classification, ordination and discrimination tests.
(ii) Univariate methods are used to determine levels of disturbance or
'stress' at given sites.
(iii) Correlation of (i) and (ii) above are tested against measured
pollution levels.
(iv) Experimental investigations are made testing cause and effect
relationships .
5.5.1 Discrimination between sites
Two basic techniques are used to discriminate between sites, namely
ordination and classification. An ordination attempts to present a picture of the
relationships between samples in terms of their similarity in species abundances
or biomass. In the, usually 2 -dimensional, picture the relative distance apart
of any pair of samples reflects their relative dissimilarity. Cluster analysis,
by contrast, forms groups of samples where samples within groups have more
similarities than those in separate groups. The methods are complementary and are
often plotted together to indicate the degree of similarity in the groupings
obtained.
Before using multivariate methods the data are usually subjected to
transformations which change the dominance weighing of species. The most widely
used transformation is that of Iog 10 (n + 1) but here a less stringent
transformation JJ is used. To illustrate the degree of stringency of a
transformation, for 100 individuals per sampling unit the transformed value
becomes :
yioo - 10, yyioo - 3.16 io glo 100 - 2
There are a large number of different ordination methods and most are
generally available as packages on main- frame computers. Some of the most widely
used are Principal Components Analysis (PCA) , Principal Coordinates Analysis
(PCoA), Multi-dimensional Scaling (MDS) and Reciprocal Averaging (RA) and its
variants such as Detrended Correspondence Analysis (DECORANA) . Clarke and Green
(1988) and Warwick and Clarke (1991) give a description of these methods in
relation to their application to pollution studies, and should be consulted for
further details. In general the preference by practicing ecologists is to use MDS
or DECORANA rather than the computationally simpler PCA and PCoA.
Clustering methods either fuse similar stations into larger and larger
groups, so-called agglomerative methods, or divide one group into smaller and
smaller dissimilar groups, so-called divisive methods. There is a wide choice of
indices on which to cluster. There is much merit in attempting to standardize to
a single method and thus here use of the Bray-Curtis coefficient (Bray and Curtis,
1957) followed by an hierarchical, agglomerative method employing group- average
linking and the results displayed as a dendrogram are recommended (Gray et al. .
1988).
There are statistical methods to determine the significance of differences
between replicated community samples in either time or space, e.g. the
simulation/permutation test ANOSIM (Clarke and Green, 1988; Clarke, 1990).
Likewise methods have been developed to examine the species that distinguish
between the site groups such as TWINSPAN (within the DECORANA package) or SIMPER
a programme developed for use with MDS analyses by Carr and Clarke (unpubl.) of
the Plymouth Marine Laboratory, U.K.
The range of multivariate analysis techniques available is large. Thus by
changing various aspects of the technique, a wide range of results can be
obtained. There is therefore, much merit in trying to standardise the techniques
used so that comparisons between localities of widely differing geographical
28
regions and pollutants can be made. To this end, suites of standardised analyses
methods are to be preferred. Many authors use DECORANA and TWINS PAN and/or a
recently developed, powerful and easy to use PC based programme, PRIMER by Carr
and Clarke of the Plymouth Marine Laboratory, Prospect Place, West Hoe, Plymouth
PL1 3DH.
Figure 12 shows sites along a heavily polluted gradient in the
Frierfjord/Langesundfjord area of Norway, hereafter called Frierfjord, (the
chemical gradient is described in Abdullah and Steffanek, 1988). Figure 13 shows
the multivariate analyses, classification and a MDS ordination. On both analyses,
sites A, E, and G are clearly separated from each other and from the group B, C
and D. Figure 14 shows data for DECORANA and RA showing similar findings. Thus
sites are separated but there is no indication of whether or not this separation
can be related to pollution.
Figure 15 shows classification and ordination analyses of macrobenthos
sampled around an oil platform in the North Sea in 1987, (Gray e al. . 1990).
There are four clear groupings of sites A, B, C, and D. Again site groups are
separated clearly but whether or not this is due to pollution can only be
ascertained from other analyses (see later).
Bellan and Bourcier (1990) show the application of similar techniques to
the benthos near Marseille suffering from organic enrichment from sewage waste and
these authors should be consulted for details of effects relevant to Mediterranean
fauna. Likewise, Zenetos and Papathanassiou (1989) have used the PRIMER programme
and multivariate analyses to show effect of tannery waste on the benthic fauna in
the Aegean Sea.
5.5.2 Univariate methods determining levels of disturbance
The multivariate methods simply separate groups of sites that have similar
faunal groupings and do not give any indication of the underlying causes of
species differences between sites. In order to investigate whether or not
pollution effects are involved, univariate methods are used. These methods fall
into two categories, those which must refer to unpolluted control sites and those
which can determine whether or not an individual site is affected.
5.5.2.1 Between site comparisons
Species, abundance, biomass. The simplest method to compare sites is to
plot the number of species, abundance and biomass - SAB (Pearson and Rosenberg,
1978) and ratios of B/A and A/S. Figure 16 shows plots for sites A to G using the
earlier data set, Figure 13. The plots show that sites A and G have highest B/A
and lowest A/S ratios suggesting that these sites are relatively undisturbed
whereas sites B, C, D and E are intermediate.
Diversity. It has long been a tradition in pollution monitoring that
diversity indices are used to assess whether or not particular sites are polluted.
This probably arose because pollution control authorities were often engineers who
were not familiar with biological complexity and simply wanted an index which
integrated both number of species and individuals per species. Diversity indices
are supposedly high at unpolluted sites and low at polluted sites. However, in
general, statistically significant reductions in a diversity index are accompanied
by such profound changes in the biological system studied that the method is a
rather poor indicator of pollution- induced change (Bayne ?t al. . 1988).
There is a wide variety of diversity indices in common usage, all of which
are highly correlated between themselves. The most widely used diversity index
is undoubtedly the Shannon-Wiener index, which is recommended here:
s
H' - - 2 Pi log P!
where p i - n^N, s - total number of species, N - total number of individuals n A
- number of i the species from 1 to s, (Shannon and Weaver, 1949).
The base of the logarithm used can vary. Some people prefer to use the base
29
Figure 12. Sites along a gradient of pollution at Frierfjord/ Langesundf jord,
outer Oslofjord, Norway, at which analyses of benthic communities
shown in Figs 13 and 14 were conducted [see Gray et al. (1988) for
details]
30
too
u
u
\ t
k
H
n
p
i
p
:
3
P
D
D
3
:
^f
G
a
pec
90
ao
70
60
50
40
30
1
I
i
i
I
I
c
r
i
JT
W
L.
abundance
D e C
'V.
w
biomaaa
oc
a o
b
iOQ AAAAEEEEGGGGBBBCC8CCDOOD
90
80
70
60
50
40
30
20
A* **
AA
V
Q
s c c
D
Figure 13. Multivariate analyses of benthic communities at Frierfjord/
Langesundfjord. (Left) Classification analysis showing dendrogram for
group-average clustering of Bray-Curtis dissimilarities (y-axis)
between 24 field macrofauna samples (x-axis), consisting of 4
replicate grabs at each of sites A to E and G. (a) Species abundance
data after JJ transformation (b) , Species biomass data after JJ
transformation. (Right) Multidimensional scaling (MDS) ordination
based on Bray-Curtis similarities between grab samples, (a) JJ
abundance (b) JJ biomass (c) raw abundance (d) raw biomass (e) JJ
abundance sites B to D only (f) JJ biomass sites B to D only (from
Gray ejt al. . 1988) J
2, that of the original formulation, but other bases e.g. log., or Iog 10 are
equally valid. Care should be taken however, in comparing diversity indices in
that the base used is both stated and similar.
Frontier (1985) has given an extensive review of diversity and related
properties in aquatic ecosystems and should be consulted for a detailed discussion
of this topic.
Another facet of the diversity concept which is not covered by the index
and which is often quoted is evenness (Pielou, 1966):
31
(a)
I
OCA I
A
A
b
OCA!
A A
C B
C
C
300
OCA1
(C)
A
A
Figure 14. Detrended Correspondence Analysis (DECORANA) ordination of data from
sites in Fig. 12 (a) no transformation (b) JJ transformation (c)
reciprocal Averaging (RA) ordination of data from sites in Fig. 12
(from Gray t al. . 1988)
J - H'/H (max)
In terms of the above equation H (max) - Iog 2 s
For the Frierfjord data, plots of diversity and evenness (Figure 16b) show
that site E had highest diversity closely followed by site A, whereas B, C and D
were lowest. Figure 17 shows data for the Statfjord oilfield in the North Sea
with low diversity near the platform at high oil levels in the sediments, and for
a mine waste site in Norway, again with low diversity near the outfall.
Although Hill (1973) calculates a range of 9 different diversity indices
ranging from the number of species to an index close to evenness there is litle
point in calculating them all as there is strong correlation between them all and
the Shannon-Wiener index is as good as any other index.
32
(a)
10
20
30
5 50 4
en
en
C 60 .
80
90
100
^1
,393427262322283324 4 93219 2 6 IB 13 13 14 1Z 29,38 33325363^2120 7 B 17 jB 11 10 8 ^3730
B C A
Figure 15. a) Classification analysis of benthic data from Ekofisk oilfield,
North Sea in 1987 (7 transformed data with Bray-Curtis
dissimilarities). Numbers refer to site numbers and letters to groups
of stations separated at 62% dissimilarity, except for group D where
two sites are separated at 50%. b) MDS ordination of the same data
with superimposed groupings from the classification analysis in (a).
(From Gray et al. . 1990)
33
Ax 100
/A
IS
A/S
/t/A
(c)
0-4
Figure 16. Plots for sites A to G from Figure 12 (a) total number of species (S),
mean abundance per 0.1 m 2 (A), mean biomass in mg per 0.1 m 2 , (b)
ratios of B/A and A/S across sites, (c) Diversity (H') and evenness
(J) across sites. (For details see Gray et al. . 1988)
(a)
J)
S
5 2
3
Q
L 1 j I I I
10 100 1000 10000 2 4 6 8 10 12 14
THC (p.p.m.) % Ti0 2
Figure 17. Relationship between species diversity and total oil concentration in
ppm for (a) Statfjord oilfield, Norwegian sector and (b) diversity and
Ti0 2 content of sediment at site of mine waste discharge in
Jossingfjord, Norway (from Gray, 1989)
34
5.5.2.2 Determination of disturbance at individual sites
Two methods have been suggested for assessing whether or not individual
sites are disturbed. The first is to plot the number of individuals among species
in geometric classes of abundance (Gray and Pearson, 1982). An undisturbed site
will have a steep curve, cover few geometric classes and have an abundance of rare
species represented by one specimen. A disturbed site will have a shallow curve,
which is often disjointed, cover many geometric classes and have few rare species.
Figure 18a shows a plot for the Frierfjord data showing site A is undisturbed,
whereas C and E are disturbed and B and D are intermediate.
The second method is the abundance, biomass comparison (ABC) method of
Warwick (1986) and Warwick et al. (1987) using 'k' dominance plots. Here the
cumulative dominance in terms of abundance and biomass are plotted against a
logarithmic scale of species rank. Figure 18b shows data for the Frierfjord sites
where at sites A and G the biomass curves lie above the abundance curves
(undisturbed) . At sites C and D the biomass curve is well below the abundance
curve indicating moderate to gross disturbance, whereas sites B and E are
intermediate .
6.5.2.3 Correlation with anthropogenic inputs
Having some indication that site differences are due to disturbance it is
important to ascertain whether this is caused by contaminant loadings or not. The
assumption is that both environmental and chemical data have been recorded at the
sites where faunal samples were taken. In the case of the Frierfjord example used
above, replicate cores were taken and a suite of PAH and metal data were analyzed.
For the Ekofisk oilfield data (Figure 15) barium is an excellent indicator of
anthropogenic activity as it is used in both oil-based and water-based drilling
muds. Thus the barium content in the sediment will be used.
For the Frierfjord data using univariate methods (number of taxa, biomass,
abundance, diversity etc.) mean biomass was particularly low at sites C and D and
sites A and G had the highest B/A and lowest A/S . This suggests that sites A and
G are relatively undisturbed whereas C, D and E are disturbed. Diversity (H r )
showed that A and E had highest values whereas B, C, D and G had similar values.
The ABC plots showed A and G as undisturbed with B and E moderately disturbed and
C and D moderately to grossly disturbed. With the exception of the diversity data
where the pattern for stations E and G differs from the other analyses, all
methods suggest that sites A and G are undisturbed with B, C, D and E being
disturbed.
A PCA analysis was done on the Frierfjord chemical data and significant
differences between sites were found, (Gray et al . . 1988). Site A had lowest
heavy metal loadings, G the highest and B, C, D and E intermediate. The measured
environmental and pollution variables were then superimposed upon the MDS analyses
of faunal data (Figure 19).
Neither sediment type (b) nor PAH (d) correlate with the faunal groupings.
The B, C, D site group has sediments with both the largest and smallest particles.
The E, G cluster has both high and intermediate levels for both metals and PAH.
However, depth (a) separates the clusters well and increases in a uniform manner
from left to right. Thus although there is possibly an effect of heavy metals the
simplest explanation is that depth plays an overriding role on the species
composition of the macrofauna. The site closest to the pollution source, G,
showed at most slight indication of organic enrichment. These results are
important as they illustrate clearly that interpretation of effects of pollutants
on marine benthos must take into account the most significant environmental
variables, (depth and grain size) otherwise one might risk attributing effects of
pollution (heavy metals) where in fact perhaps only depth is significant. These
data again illustrate the necessity in sample design of making sure that
'nuisance 1 variables are adequately covered. Here a revised sampling strategy in
the light of the above findings would be to sample at similar depths in the
future .
In the case of the Ekofisk oilfield, the data are from similar depths and
grain size. Again diversity was not found to be a particularly useful variable
as only the grossly polluted sites had lower values than the unpolluted sites.
(a)
. N-N ....
2 Of:
\
MI iv v vi vii VIM
I II HI IV V VI VII VIII \X * K\
NUMBER OF INDIVIDUALS PER SPECIES 1X1 GEOMETRIC CLASSES)
(b)
SPECIES RANK
Figure 18. Detection of disturbance at individual sites, (a) Plots of number of
species against number of species per individual in geometric classes
at sites A to E and G from Figure 12. (class intervals I - 1
individual per species, II - 2-3 individuals per species (III - 4-7
individuals per species etc), (b) Abundance biomass comparison curves
(ABC) for sites A to E and G in Figure 12. Abundance squares and
solid line, biomass crosses and broken line based on the totals from
4 replicates at each site (For details see Gray et al. . 1988)
36
I
I
r 11 !!
6bo
o"
o
fcb
Figure 19. MDS ordination of the 24 field macrofaunal samples from sites A to E
and G in Figure 12. The data are shown with superimposed symbols on
the original faunal groupings in linear dimensions proportional to the
values of the selected environmental variables, (a) water depth,
smallest symbol 22m largest 113m. (b) median particle diameter of
sediment, smallest 7.8 /im largest 16.5 /im. (c) metal concentration in
sediment mean principal component score from PC analysis representing
an average of Cu, Zn, Pb, Ni, Cr, Cd levels, smallest -2.9 largest +
3.2 (d) total PAH concentrations in sediment, smallest 4.4 pg g" 1
largest 14.8 /ig g" 1 (Gray et al. . 1988)
Figure 20 shows site groupings from the multivariate analyses and the sediment
chemistry for the station groupings obtained by the multivariate analyses.
Clearly the separation into the non-polluted group A and the just-affected group
B can be related to barium content, with a mean of 6.3 at A and 7.6 at B.
However, there is no separation between B and C and D on the barium content, all
having statistically similar values around 7.6 - 7.8. Yet D has the highest total
hydrocarbon content (6.25) which is statistically higher than the other three
groups A, B and C. C (mean 7%) is intermediate between both A and B (4%) and D
(12%) in the percentage mud content. As the percentage mud probably represents
the amount of drilling material deposited, a pollution gradient is reflected from
A-D. Thus the Ekofisk example shows clearly how multivariate data can be combined
with environmental data and statistical analyses to show the influence of
pollutants on benthic communities.
5.5.2.4 Analyses of higher taxa only
One of the most interesting aspects of recent research on effects of
pollutants on benthic communities is that effects similar to those shown above can
be found when higher taxal levels than the species are used (Warwick, 1988).
Using data from the Frierfjord macrofaunal studies Warwick mis able to show that
analyses done at the level of families of macrofauna gave equally good separation
of the faunal groups as that obtained using species (Figure 21). The level of
phyla gave less tight clusters but still separated the major station groupings.
Similar results were obtained from analyses of the Ekofisk data and for meiofaunal
analyses of the Frierfjord data (Heip t al. . 1988).
These results suggest that it might be possible to greatly simplify
pollution monitoring work by i J _ng only to the level of family (and possibly
only to phylum). Whilst more substantiation of these results are needed, this
37
(a)
30.000 y 9 ,
M
T
IS
"
12
{
* 6
: i
i i
3
.
A B C
GROUPS FROM CLUSTERING
a. s
B
I 7 - 8
j I ;;
7
.
s
6.S
6
; i
10
B
I
A B C
GROUPS FROM CLUSTERING
A B C
GROUPS FROM CLUSTERING
Figure 20. (a) Plots of the site groupings from the multivariate analyses of the
Ekofisk macrofauna (Figure 15). Ekofisk 2/4B & K are newly bored holes
whereas the Ekofisk Centre is the original oilfield site. The data
show 4 groupings with A sites being unpolluted and D being grossly
polluted, (b) Plots of selected environmental variables for the site
groupings from Figure 15 (Gray et al. . 1990)
(a)
'
"A
o c
A A
A 'Jb
00 .l
*
c
C
A A P
c
A A A A 00 B
1
V 1 '
B
' \* '"
e o o /,". =
A A
o B |
c
<Po
a >c
* f
tf
v.
J" 1
'
a
o
'I
to
Figure 21. Use of higher taxa in pollution studies, (a) MDS analysis comparing
the original species analysis with analyses for families and phyla for
abundance data, (b) as in (a) but for biomass data, (c) ABC plots for
data using family data (cf Figure 18b for species) [see Warwick (1988)
for details]
- 39 -
points the way to possible large savings in costs of benthic pollution surveys*
It also suggests that perhaps statistical techniques for analyzing benthic data
sets need more research, as ecological theory would suggest that it is species
that should best reflect environmental gradients rather than families or phyla.
6* REFERENCES
Abdullah, M.I* and I. Steffanek, 1988. The GEEP Workshop: Trace metal analyses.
Mar . Ecol . Proor . Ser . . 46:27-30
Abernethy, S. and D. Mackay, 1986. Organic contamination structure-property
toxicity relationships for aquatic organisms. Technology Transfer
Conference Part B Water Quality Research, Ministry of the Environment,
Ontario, Canada. Corp. Resour. Div. pp. 113-137
Addison, R. , 1984. Hepatic mixed function oxidase (MFO) induction in fish as a
possible biological monitoring system. In: Contaminant effects on
fisheries, edited by V.W. Cairns, P.V. Hodgson and J.O. Nriagu, Toronto,
Wiley, pp. 51-60
Addison, R. and A.J. Edwards, 1988. Hepatic microsomal mono-oxygenase activity
in flounder Platichthys f Jesus from polluted sites in Langesundf jord and
from mesocosms experimentally dosed with diesel oil and copper.
Mar . Ecol . Proor . Ser . , 46:51-54
Augier, H., 1982. Inventaire et classification des bioc&noses marines benthiques
de la MSditerran&e. Conseil d 'Europe, Strasbourg. 58 p.
Bailey, J.H., A. Nelson-Smith and E.W. Knight-Jones, 1967. Some methods for
transects across steep rocks and channels. Underwater Association Report,
1966-67, pp. 107-111
Bayne, B.L., 1980. Physiological measurements of stress. In; Biological
effects of marine pollutants and the problems of monitoring, edited by A.D.
Mclntyre and J.B. Pearce. Rapp . P . -V . R6un . CIEM , 179:56-61
Bayne, B.L. and C.M. Worral, 1980. Growth and production of mussels Mvtilus
edulis from two populations. Mar . Ecol . Proor . Ser . . 3:317-328
Bayne, B.L., K.R. Clarke and J.S. Gray (Eds.), 1988. Biological Effects of
pollution: Results of a Practical Workshop. Mar . Ecol . Proor . Ser * / 46:1-278
Bellan, G., 1981. Manual of methods in aquatic environment research. Part 7.
Selected bioassays for the Mediterranean. FAQ Fish. Tech. Pap. , (208) :31 p.
Bellan, G., 1985. Effects of pollution and man-made modifications on marine
benthic communities in the Mediterranean: A Review. In: Mediterranean
marine ecosystems edited by M. Moraitou-Apostolopoulou and V. Kiortsis, New
York, Plenum Publishing Co., pp. 163-194
Bellan, G. and M. Bourcier, 1988. Etude & long terme (1976-1986) des peuplements
de substrats meubles au large de I'&nissaire de Marseille Cortiou.
Rapp . P . -V . R6un . CIESM . 31(2):168
Bellan, G. and M. Bourcier, 1990. Les enseignements d'une 6tude sur dix ans
(1976-1986) des peuplements de substrata meubles au large d'un emissaire
d'eaux uses: Marseille-Cortiou. Cah.Biol.Mar. , 31:225-249
Benovi6, A., J. Dubravko and A. Bender, 1978. Enigmatic changes in the
hydromedusan fauna of the northern Adriatic sea. Nature fLond.l. 326:597-
600
Boynton, W.R. et al. , 1981. The influence of water circulation rate on in situ
measurements of benthic community respiration. Mar.Biol. , 65:185-190
Brandt, A. von, 1981. Fish catching methods of the world. West Byfleet, Surrey,
Fishing News (Books), 240 p.
- 40 -
Brey, T., 1986. Estimation of annual P/B-ratio and production of marine banthic
invertebrates from length- frequency data. Ophelia SUDD!, . 4s 45-54
Brey, J.R. and J.T. Curtis, 1957. An ordination of the upland forest communities
of Southern Wisconsin. Ecol.Monoor. . 27x325-349
Brey, T., H. Rumohr and S. Ankar, 1988. The energy content of macrobenthic
invertebrates. General conversion factors from weight to energy.
J.Exp.Mar.Biol.Ecol. . 117:271-278
Bucke, D. and B. Watermann, 1988. Effects of pollutants on fish. In: Pollution
of the North sea. An Assessment, edited by W. Salomons, B.L. Bayne, E.K.
Duursma and U. Forstner, Springer-Verlag, pp. 612-623
Capuzzo, J.M. and B.A. Lancaster, 1985. Zooplankton population responses to
industrial wastes discharged at deepwater dumpsite-106. In; Deep sea waste
disposal, edited by D.A. Hester et al. . New York, Wiley and Sons, pp.
209-226
Chapman, P.M. and E.R. Long, 1983. The use of bioassays as part of a
comprehensive approach to marine pollution assessment. Mar. Pollut* Bull. .
14:81-84
Christie, H., 1983. Use of video in remote studies of rocky subtidal community
interactions. Sarsia, 68:191-194
Christie, H., R.A. Evans and O.K. Sandnes, 1985. Methods for in situ subtidal
hard bottom studies. In: Underwater photography and television for
scientists, edited by J.D. George., G.I. Lythgoe and J.N. Lythgoe, Oxford
Scientific Publications, pp. 37-48
Claisse, D., 1989. Chemical contamination of French coasts. The results of a
ten years mussel watch. Mar. Pollut. Bull. . 20:523-528
Clarke, K.R., 1990. Comparisons of dominance curves. J.Exp.Mar.Biol.Ecol. ,
138:143-157
Clarke, K.R* and R.H. Green, 1988. Statistical design and analysis for a
'Biological Effects' study. Mar . Ecol . Procrr . Ser . . 46:213-226
Connor, P.M., 1972. Acute toxicity of heavy metals to some marine larvae.
Mar. Pollut. Bull.. 3:190-192
Craib, J.S., 1965. A sampler for taking short undisturbed cores. J. Con a. C I EM. .
30:34-39
Davenport, J., 1982. Oil and plankton ecosystems. Phi 1 . Trans . R . Soc . Lond . ,
B297, pp. 369-384
Davies, J.M. g al. , 1984. Environmental effects of oil-based mud cuttings.
Mar. Pollut. Bull. , 15:363-370
De Jonge, U.N. and L.A. Bouwman, 1977. A simple density separation technique for
quantitative isolation of meiobenthos using colloidal silver Ludox-TM,
Mar.Biol., 42:143-148
Deneer, J.W. g al. . 1988. The joint acute toxicity to Daohnia maona of
industrial organic chemicals at low concentrations. Aouat.Toxicol. ,
12:33-38
Duineveld, G.C.A. and H.J. Witte, 1987. Report on an inter-calibration exercise
on methods for determining ashfree dry weight of macrozoobenthos. ICES
Doc. C.M. 1987/L:39
Eagle, R.A. and E.I.S. Rees, 1973. Indicator species - a case for caution.
Mar. Pollut. Bull.. 4:25
41
Eleftheriou, A. and N.A. Holme, 1984. Macrofauna Techniques, p. 140-216. In:
Methods for the study of marine benthos, edited by N.A. Holme and A.D.
Mclntyre, Blackwell, 387 p.
Elliott, J.M., 1971. Some Methods for the Statistical Analysis of Samples of
Benthic Invertebrates. FBA Scientific Publication, 144 pp 25, Freshwater
Biological Association, Arable side
Field, J.G., K.R. Clarke and R.M. Warwick, 1982. A practical strategy for
analysing multispecies distribution patterns. Mar . Ecol . Progr . Ser . . 8:37-52
Frontier, S., 1985. Diversity and structure in aquatic ecosystems. Oceanogr.Mar.
Biol.Ann.Rev.. 23:253-312
Gaffney, P.M. and W.J. Diehl, 1986. Growth, condition and specific dynamic action
in the mussel Mytilus edulis recovering from starvation. Mar . Biol. .
93:401-409
Gamble, J.C., 1984. Diving. In: Methods for the study of marine benthos, edited
by N.A. Holme and A.D. Mclntyre, Oxford, Blackwell, pp. 99-139
GESAMP, 1990. The State of the Marine Environment. Rep.Stud.GESAMP (39). Ill p.
Goldberg, E.D., 1975. The mussel watch. A first step in global marine
monitoring. Mar. Pollut. Bull. . 6:111-113
Grassle, J.F. and J.P. Grassle, 1976a. Opportunistic life histories and genetic
systems in marine benthic polychaetes. J. Mar. Res. . 32:253-284
, 1976b. Sibling species in the marine pollution indicator Capltella
(Polychaeta). Science (Wash.). 192:567-569
, 1977. Temporal adaptations in sibling species of Capitella. In:
Ecology of marine benthos, edited by B.C. Coull, Belle W. Baruch Library
in Marine Science 6, S. Carolina, University Press, Columbia, pp. 179-189
Gray, J.S., 1981. The Ecology of Marine Sediments. In: Modern biology 2.
Cambridge Studies, Cambridge Univ. Press, 181 p.
, 1989. Effects of environmental stress on species -rich assemblages.
Eiol.J.Linn.Soc. . 37:19-32
Gray, J.S. and T.H. Pearson, 1982. Objective selection of sensitive species
indicative of pollution- induced change in benthic communities. 1:
Comparative methodology. Mar . Ecol . Progr . Ser . . 9:112-119
Gray, J.S. et al. . 1988. Analysis of community attributes of the benthic
macrofauna of Frierf jord/Langesund and in a mesocosm experiment.
Mar. Ecol. Progr. Ser. . 46:151-165
Gray, J.S. t al. . 1990. Detection of initial effects of pollution on marine
benthos: an example from the Ekofisk and Eldfisk oilfields, North Sea.
Mar. Ecol. Progr. Ser. . 66:285-299
Grice, G.D. and M.R. Reeve, 1982. Marine mesocosms: biological and chemical
research in experimental ecosystems. New York, Springer, 430 p.
Help, C. et al. . 1988. Analysis of community attributes of the benthic me io fauna
of Frierf jord/Langesund. Mar . Ecol . Progr . Ser . . 46:171-180
Hessler, R.R. and P. A. Jumars, 1974. Abyssal community analysis from replicate
box cores in the central North Pacific. Deep Sea Res . . 21:185-209
Higgins, R.P. and H. Thiel, 1988. Introduction to the Study of Meiofauna.
Washington D.C., Smithsonian Institution Press, 488 p.
Hill, M.O., 1973. Diversity and evenness: A unifying notation and its
consequences. Ecology. 54:427-432
42
Holme, N.A., 1984. Photography and Television. In: Methods for the study of
marine benthos, edited by N.A. Holme and A.D. Mclntyre, Oxford, Blackwell,
pp. 66-98
Holme, N.A. and A.D. Mclntyre, 1984. Methods for the Study of Marine
Benthos. 2nd Edition. Oxford, Blackwell, 387 p.
Hulings, C.N. and J.S. Gray, 1971. A manual for the study of me io fauna.
Smithsonian Contrib.Zool. . (78):84 p.
ICES, 1978. On the feasibility of effects monitoring. Coop. Res. Rep. 75.
International Council for the Exploration of the Sea, Denmark, 42 p.
ICES, 1990. Soft bottom macro fauna: Collection and treatment of samples.
Techniques in Marine Environmental Sciences. No. 8. Compiled by H. Rumohr.
International Council for the Exploration of the Sea, Denmark, 18 p.
Jensen, A., 1980. The use of phytoplankton cage culture for in situ monitoring
of marine pollution. Rapp.P. -V.R<un.CIEM. 179:306-309
Kobayashi, N. , 1971. Fertilised sea urchin eggs as an indicatory material for
marine pollution bioassay, preliminary experiments. Publ Seto
Mar.Biol.Lab. . 18:379-406
Koehn, R. and B.L. Bayne, 1989. Towards a physiological and genetical
understanding of the energetics of the stress response.
Biol.Journ.Linn.Soc. . 1989. 37:157-171
Kuiper, J. and J.C. Gamble, 1988. Between test tubes and North Sea: Mesocosms.
In: Pollution of the North sea. An Assessment, edited by W. Salomons, B.L.
Bayne, E.K. Duursma and U. Forstner, New York, Springer-Verlag, pp. 638-654
Larkum, A.W.D., E.A. Drew and R.N. Crossett, 1967. The vertical distribution of
attached marine algae in Malta. J. Ecology. 55:361-371
Lee, R. et al. . 1980. Biochemical techniques for monitoring biological effects
of pollution. Rapp.P.-V.R^un.CIEM. 179:48-55
Lincoln, J.R. and L. Sheals, 1979. Invertebrate animals (collection and
preservation). Cambridge, British Museum. 150 p.
Loeb, S.C., 1981. An in situ method for measuring the primary productivity and
standing crop of the epilithic periphyton community in benthic systems.
Limnol.Oceanogr. . 26:394-399
Long, E.R. and P.M. Chapman, 1985. A sediment quality triad: Measure of sediment
contamination, toxicity an
Mar. Pollut. Bull. . 16:405-415
contamination, toxicity and infaunal composition in Puget Sound.
Longwell, A. Crosby and J.B. Hughes, 1980. Cytological, cytogenetic and
developmental state of Atlantic mackerel eggs from sea surface waters of
the New York Bight, and prospects for biological effects monitoring with
ichthyoplankton. Rapp.P. -V.Reun.CIEM. 179:275-291
LundAlv, T., 1971. Quantitative studies on rocky-bottom biocoenoses by
underwater photogrammetry : A methodological study. Thalassia Jugosl..
7:201-208 ,-.,
Mclntyre, A.D., 1971. Deficiency of gravity corers for sampling meiobenthos and
sediment. Nature (Lond.^. 231:260
Mclntyre, A.D. and J.B. Pearce, 1980. Biological effects of marine pollution and
the problems of monitoring. Rapp . P . - V . R6un . CIEM . 179:1-346
Mclntyre, A.D. and R.M. Warwick, 1984. Mesofauna techniques, pp 217-244. IB:
Methods for the study of marine benthos, edited by N.A. Holme and A.D.
Mclntyre, Oxford, Blackwell, 387 p.
43
Main, J. and G.I. Sangster, 1978. The value of direct observation techniques by
divers in fishing gear research. Scott . Fish . Res . Rep . . (12): 15 p.
Miller, D.C., 1980. Some applications of locomotor response in pollution effects
monitoring. Rapp.P. -V.R6un.CIEM. 179:154-161
Mirza, F.B. and J.S. Gray, 1981. The fauna of benthic sediments from the
organically enriched Oslofjord, Norway, J . Ext) . Mar . Biol . Ecol . . 54:181-207
Moore, M.N. , 1988. Cytochemical responses of the lysosomal system and
NADPH-ferrihemoprotein reductase in molluscan digestive cells to
environmental and experimental exposure to xenobiotics.
Mar . Ecol . Progr . Ser . . 46:81-89
Moore, M.N. and D.M. Love, 1985. Cytological and cytochemical measurements. In:
The effects of stress and pollution on marine animals, edited by B.L. Bayne
et al. . New York, Praeger, pp. 46-74
Moore, H.B. and R.G. Neill, 1930. An instrument for sampling marine muds.
J.Mar.Biol.Assoc.UK.. 16:589-594
Murphy, L.S., E.M. Haugen and J.F. Brown, 1983. Phytoplankton: Comparison of
laboratory bioassay and field measurements. In: Industrial and sewage
wastes in the ocean, edited by I.W. Duedall et al. . New York, Wiley and
Son, pp. 219-233
NOAA, 1979. Diving Manual. National Oceanographic and Atmospheric Administration
Manned Undersea Science and Technology Office. US Government Printing
Office, Washington D.C
Olla, B.L., W.H. Pearson and A. L. Studholme, 1980a. Applicability of behavioural
measure, in environmental stress assessments. Rapp.P. -V.Reun.CIEM.
179:162-173
Olla, B.L. et al. . 1980b. The role of behaviour in marine pollution monitoring.
Rapp. P.- V.Reun.CIEM. 179:174-181
Pearson, T.H. and R. Rosenberg, 1978. Macrobenthic succession in relation to
organic enrichment and pollution of the marine environment.
Oceanogr. Mar. Biol. . 16:229-311
Perry, D.M. , J.S. Weiss and P. Weiss, 1988. Cytogenetic effects of methylmercury
in embryos of killifish, Fundulus heteroclitus. Arch. Environ.
Cont.Toxicol.. 17:569-574
Picard, J., 1965. Recherches qualitatives sur les biocenoses marines des
substrats meubles de la region marseillaise . Reel . Trav . Stn . Mar . Endoume .
52 (Bull.36):l-160
Pielou, E.G. , 1966. The measurement of diversity in different types of biological
collections. J .Theor.Biol. . 13:131-144
Portmann, J.E., 1976. Manual of methods in aquatic environment research. Part 2.
Guidelines for the use of biological accumulators in marine pollution
monitoring. FAQ Fish. Tech. Pap. . (150): 76 p.
Price, J.H., D.E.G. Irvine and W.F. Farnham (Eds.), 1980. The Shore Environment
vol. 1. Methods, Systematics Association Special Volume 17(a). London,
Academic Press, 321 p.
Pravdic, V., 1981. GESAMP, The First Dozen Years. UNEP, 29 p.
Reiersen, L.O. , J.S. Gray, K.H. Palmork and R. Lange, 1989. Monitoring in the
vicinity of oil and gas platforms; results from the Norwegian sector of the
North Sea and recommended methods for forthcoming surveillance In: Drilling
wastes, edited by F.R. Engelhardt, J.P. Ray and A.H. Gillam. London,
Elsevier, pp. 91-117
- 44 -
Reineck, H.E., 1958. Kastengreifer und Lotrohre 'Schnepfe', Cerate zur Entnahme
ungestorter, orientierter Meeresgrund proben. Senckenberoiana lethaea .
39:42-48; 54-56
Reish, D.J., 1970. The effects of varying concentrations of nutrients, chlorinity
and dissolved oxygen on polychaetous annelids. Water Res . , 4x721-735
Reish, D.J. and J.L. Barnard, 1960. Field toxicity tests in marine waters
utilising the polychaetous annelid Caoitella caoitata (Fabricius) . Pacific
Naturalist. 1:1-8
Reish, D.J. and Oshida, 1986. Manual of methods in aquatic environment research.
Part 10. Short term static bioassays. FAQ Fish. Tech. Pap. , (247) :62 p.
Revelante, N., N. Gilmartin and N. Smodlaka, 1985. The effects of the Po river
induced eutrophication on the distribution and community structure of
ciliated Protozoan and Micrometazoan populations in the Northern Adriatic
sea. J. Plankton Res. . 7:461-471
Rhoads, D.C. and J.D. Germano, 1982. Characterization of organ ism- sediment
relations using sediment profile imaging: An efficient method of Remote
Ecological Monitoring of the Seafloor (REMOTS system). Mar . Ecol . Procrr . Ser . .
8:115-128.
Rumohr, H., T. Brey and S. Ankar, 1987. A compilation of biometric conversion
factors for benthic invertebrates of the Baltic Sea. Bait . Mar . Biol . Publ . .
9:56 p.
Sanders, H.L., R.R. Hessler and G.R. Hampson, 1965. An introduction to the study
of deep sea faunal assemblages along the Gay Head - Bermuda transect. Deep
Sea Res.. 12:845-867
Shannon, C.E. and W. Weaver, 1949. The mathematical theory of communication.
Urbana, Illinois, University of Illinois Press, 117 p.
Sinderman, C.J., 1983. An examination of some relationships between pollution
and disease. Rapp . P . -V . R6un . CIEM . 182:37-43
Sinderman, C.J. et al. . 1980. The role and value of pathobiology in pollution
effects monitoring programmes. Rapp . P . -V . Run . CIEM , 79:135-151
Stebbing, A.R.D., 1985. The use of hydroids in toxicology. In: The effects of
stress and pollution on marine animals, edited by B.L. Bayne et al. , New
York, Praeger, pp. 267-297
Steedman, H.F., 1976. Zooplankton fixation and preservation. Paris, UNESCO
Press, 350 p.
Stegeman, J.J., 1980. Mixed-Function oxidase studies in monitoring for effects
of organic pollution. In: Biological effects of marine pollution and the
problems of monitoring, edited by A.D. Mclntyre and J.B. Pearce,
Rapp.P.-V.R6un.CIEM. 179:33-38
Stegeman, J.J., B.R. Woodin and A. Goksoyr, 1988. Cytochrome P-450 induction in
flounder. In; Biological effects of pollution, edited by B.L. Bayne, K.R.
Clarke and J.S. Gray, Mar . Ecol . Proor . Ser . . Special Volume, 46:1-278
Stirn, J. et al. . 1975. Selected biological methods for assessment of marine
pollution. In; Marine pollution and marine waste disposal, edited by E.A.
Pearson and E. de Fraja Frangipane, Oxford, Pergamon Press, pp. 307-327
Tanner, C. et; al. . 1977. A hand-operated suction sampler for the collection of
subtidal organisms. J. Fish. Res. Bd. Can. , 34:1031-1034
UNEP/FAO/IAEA, 1987a. Test of the acute lethal toxicity of pollutants to marine
fish and invertebrates. Reference method for marine pollution studies No
43. UNEP, 27 p.
- 45 -
UNEP/FAO/IABA, 1987b. Estimation of the toxicity of pollutants to marine
phytoplankton and zooplankton organisms. Reference methods for marine
pollution studies No 44. UNEP, 24 p.
UNEP/FAO/IAEA, 1987c. Comparative toxicity test of water soluble fractions of oils
and oil dispersants to marine organisms. Reference methods for marine
pollution studies No 45. UNEP, 21 p.
Underwater Association for Scientific Research Ltd., 1979. Code of practice for
scientific diving, 3rd. Edition. Natural Environment Research Council,
Swindon, 82 p.
Uhlig, G., H. Thiel and J.S. Gray, 1973. The quantitative separation of
meiofauna. HelQol.Wiss.Meeresunters. . 25:173*195
Uthe, J.F. fit al. . 1980. Selection of biochemical techniques for detection of
environmentally induced sublethal effects on organisms.
Rapp . P . -V . R6un . CIEM , 179:39-47
Venrick, E.L., 1978. Sampling strategies. In: Phytoplankton manual, edited by
A. Sournia, UNESCO Monographs on Oceanographic Methods, pp. 7-16
Viarengo, A. et al. . 1988. Integrated cellular stress indices in trace metal
contamination: Critical evaluation in a field study. Mar . Ecol . Proqr . Ser . ,
46:65-70
Warwick, R.M., 1986. A new method for detecting pollution effects on marine
macrobenthic communities. Mar.Biol. . 92:557-562
, 1988. Analysis of community attributes of the macrobenthos of
Frierf jord/Langesundf jord at taxonomic levels higher than the species.
Mar . Ecol . Procrr . Ser . . 46:167-170
Warwick, R.M. and K.R. Clarke, 1991. A comparison of some methods for analysing
changes in benthic community structure. J.Mar.Biol.Assoc.U.K. , 71:225-244
Warwick, R.M., T.H. Pearson and Ruswahyuni, 1987. Detection of pollution effects
on marine macrobenthos: Further evalution of the species/abundance method.
Mar.Biol.. 95:193-200
Wethy, D.S., 1984. Spatial pattern in barnacle settlement: Day-to-day changes
during the settlement season. J.Mar.Biol.Assoc.U.K., 64:687-698
Widdows, J., 1985. Physiological measurements. In; The effects of stress and
pollution on marine animals, edited by B.L. Bayne et al. , New York,
Praeger, 384 p.
Widdows, J. and D. Johnson, 1988. Physiological energetics of molluscs. In:
Biological effects of pollutants, edited by B.L. Bayne, K.R. Clarke and
J.S. Gray, Mar . Ecol . Proor . Ser . . Special Volume, 46:1-278
Winberg, G.G., 1971. Methods for the estimation of the production of aquatic
animals. London, New York, Academic Press, 175 p.
Winther, U. and J.S. Gray, 1985. The biology of Mva arenaria (Bivalvia) in the
eutrophic inner Oslofjord. Sarsia, 70:1-10
Woelke, C.E., 1967. Measurement of water quality with the Pacific oyster embryo
bioassay. Water Quality Criteria. 416:112-120
, 1968. Application of shellfish bioassay results to the Puget Sound
Pulp Mill pollution problem. N.Y. Sci.. 42:125-133
Woelke, C.E., 1972. Development of a receiving water quality bioassay criterion
based on the 48-hour Pacific oyster Crassostrea aiaas embryo. Washington
Department of Fisheries, Technical Report 9:1-93
- 46 -
Zenetos, A. and E. Papathanassiou, 1989. Community parameters and mult ivar late
analysis as a means of assessing the effects of tannery effluents on
macrobenthos. Mar. Pol lut. Bull.. 20:176-181
- 47 -
APPENDIX 1
Optimal allocation of samples
Samples can be allocated according to the variability of the standard error.
To take an example: the area is divided up first according to sediment types using
methods similar to those shown above and a preliminary sampling done. The total
number of animals found in this preliminary survey taking 7 replicates per station
is:
REPLICATES
Stratum
2
3
4
Station
1
2
3
4
5
6
7
8
9
10
11
12
13
14
1020
390
140
140
420
370
620
390
40
150
730
1380
1620
1850
B
1180
490
440
150
950
420
1390
430
20
140
670
1410
320
2060
1300
210
360
190
350
700
380
110
350
660
470
1190
1550
1090
2100
360
150
160
150
100
450
440
60
320
340
2710
760
2410
980
220
490
150
180
200
480
110
80
240
930
1600
1250
1520
900
310
1070
180
330
190
2600
180
20
880
370
1290
1990
220
Using a statistics package on a PC (e.g. Statgraphics) calculate the means,
standard deviation and standard errors for these data, giving:
Stratum Sample Size (n)
1
2
3
4
5
21
14
14
21
28
Mean
677.62
260.00
642.14
309.05
1162.86
s.d
504.62
218.32
654.90
381.59
686.68
TOTAL
s.e
110.12
58.35
175.03
83.27
129.77
556.54
Calculate each s.e. as a proportion of the Total s.e. and use this
proportion to calculate sampling allocation per stratum. Here it is assumed that
a total of 65 samples can be taken in the next survey.
Stratum
1
2
3
4
5
Proportion of Total s.e.
0.198
0.104
0.315
0.149
0.232
No. of Samples/Stratum
13
7
20
10
15
It is now possible to calculate how effective this sampling system has been
compared with random sampling. First an analysis of variance (using Statgraphics)
is ran on the data testing within strata variance compared with between stn
variance. This gives:
strata
48
Source of variation Sum of Squares d.f Mean Square F Ratio
Between Strata 11902293 4 2975573.3 10.27
Within Strata 26931369 93 289584.6
TOTAL 38833662 97 400347.0
The pooled standard deviation within strata s w is:
s w - 289584.6
s w - 538.13
Estimated s.e. of Y it - s(Y, t )
s( Y . t ) - s w / n n - 98
- 538. 13/ 98
- 54.36
With purely random sampling S y - s/ n
S y - 400347/ 98
- 63.91
Therefore the stratified sampling reduces the s.e. by
(63.91 - 54.36) * 100)/63.91 %
14.9 %
This is a big change and illustrates the advantages of stratified sampling.
49
APPENDIX 2
Calculation of number of replicates to be taken
Assume that 10% error is acceptable and call this proportion (O.l)D. The
number of samples that should be taken (n) is:
n - s 2 / (O.I) 2 x 2
- 100 s 2 /x 2 for a 10% accepted error.
Example: counts - 14,15,12,7,8,14,11,14,10,9,10
s 2 - 7.42 x - 11.273
First, test whether or not the distribution is random . Using the s 2 /x ratio.
If s 2 /x < 1 the distribution is REGULAR
If s 2 /x - 1 the distribution is RANDOM
If s 2 /x > 1 the distribution is AGGREGATED
(Exact tests of these distributions can be found in Elliott, 1971).
Here s 2 /x - 7.42/11.273 - 0.66 which we accept as random
n - 100 * 7.42/(11.273) 2
- 5.82 i.e. 6 samples
Most marine species however, are aggregated.
For an aggregated distribution:
Counts: 98,22,72,214,67
s 2 - 5202.8 x - 94.60
n - 100 * 5202.8/(94.6) 2
- 58 replicates
This is an enormous number of samples and is clearly impractical. So accept
a lower error estimate e.g. 20%.
n - 25 * 5202.8/(94.6) 2
- 14.53 replicates (i.e. 15)
MANUAL OF METHODS IN AQUATIC ENVIRONMENT RESEARCH
MANUEL DES METHODES DE RECHERCHE SUR L'ENVIRONNEMENT AQUATIQUE
MANUAL DE METODOS DE INVESTIGACION DEL MEDIO AMBIENTE ACUATICO
Part 1: Methods for detection, measurement and monitoring of water pollution. 1975.
FAQ Fish.Tech.Pap.. <137):238 p.
Part 2: Guidelines for the use of biological accumulators in marine pollution monitoring,
edited by J. Portmann. 1976. FAQ Fish.Tech.Pap., (150):76 p.
Part 3: Sampling and analyses of biological material (Guidelines for the FAO(GFCM)/UNEP
Joint Coordinated Project on Pollution in the Mediterranean), by M. Bernhard.
1976. FAQ Fish.Tech.Pap.. (1581:124 D.
Echantillonage et analyse du materiel biologique (Directives destinies au projet
commun coordonn6 FAO(CGPM)/PNUE sur la pollution en M6diterran6e), par M.
Bernhard. 1977. FAQ Doc.Tech.Pfiches. (1581:132 D.
Part 4: Bases for selecting biological tests to evaluate marine pollution. 1 977. FAQ Fish.
Tech. Pap.. (164):31 p.
Principes de selection des tests biologigues pour revaluation de la pollution marine.
1981. FAQ Doc.Tech.PSches, (164):32 p.
Bases para la eleccibn de ensayos bio!6gicos para evaluar la contaminacitin marina.
1981. FAQ, Doc.T6c.Pesca. (164):34 p.
Part 5: Statistical tests, by F. Moller. 1979. FAQ Fish.Tech.Pap.. (182): 131 p.
Part 6: Toxicity tests, by G.S. Ward and P.R. Parrish. 1982. FAQ Fish.Tech.Pap..
(185):23 p.
Tests de toxicite, par G.S. Ward et P.R. Parrish. 1983. FAQ Doc.Tech.Pfeches.
(185):24p.
Ensayos de toxicidad, por G.S. Ward y P.R. Parrish. 1983. FAQ. Doc.T6c.Pesca,
(185);25 p.
Part 7: Selected bioassays for the Mediterranean (Tests used by the FAO(GFCM)/UNEP
Joint Coordinated Project on Pollution in the Mediterranean), by G. Bellan. 1981.
FAQ Fish.Tech.Pap.. (208):31 p.
Bioessais s6lectionn6s pour la M6diterran6e (Tests utilises dans le projet commun
coordonne FAO(CGPM)/PNUE sur la pollution en M6diterran6e) / par G. Bellan.
1981. FAQ Doc.Tech.PSches, (208):37 p.
Part 8: Ecological assessment of pollution effects (Guidelines for the FAO(GFCM)/UNEP
Joint Coordinated Project on Pollution in the Mediterranean), by J. Stirn. 1981.
FAQ Fish.Tech.Pap.. (209):70 p.
Evaluation des modifications des 6cosyst6mes marins dues la pollution (Directives
destinies au projet commun coordonn6 FAO(CGPM)/PNUE sur la pollution en
M6diterran6e), par J. Stirn. 1982. FAQ Doc.Tech.Pfeches, (209):75 p.
Part 9: Analyses of metals and organochlorines in fish. 1983. FAQ Fish. Tech. Pao..
(21 21:33 p.
Analyse des mdtaux et des organo-chlords contenus dans les poissons. 1 983. FAQ
Doc.Tech.Pfiches. (212):35 p.
Analisis de presencia de metales y organoclorados en los peces. 1 983. FAQ. Doc.
Tdc.Pesca. (212):35p.
Part 10: Short-term static bioassays, by D.J. Reish and P.S. Oshida. 1986. FAQ Fish.
Tech.Pap.. (247):62 p.
Part 1 1 : Biological assessment of marine pollution with particular reference to benthos, by
J.S. Gray, A.D. Mclntyre and J. Stirn. 1991. FAQ Fish. Tech. Pan.. (324):49 p.