Marine biodiversity is threatened by overexploitation, physical alteration, pollution, global atmospheric change and introduction of alien species. There is, therefore, a need to acquire knowledge about patterns and processes of marine biodiversity, in order to establish effective management plans. It is only by considering genetic diversity, too often neglected by stakeholders, and the basic processes that rule biological diversity, that a given plan will have long-term success. Within this context, more-and-more refined molecular techniques allow us to obtain an important insight into natural populations by collecting information on several issues, such as gene flow, mixing of different gene pools, effective population size, inbreeding rate, genetic loss, the occurrence of recent bottlenecks, assignment of individuals to a population, the action of natural selection, gene introgression and hybridisation, taxonomic position. Data gathered can be used to provide invaluable suggestions for the management and conservation of biological marine resources.

Some main themes were addressed in the conference. The first was about species' genetic structuring, connectivity among populations and habitat fragmentation. Ferruccio Maltagliati introduced this theme providing a definition of "connectivity" and starting a discussion on the influence on dispersal of barriers and life history traits, such as the presence of pelagic larval stages. He also raised the problem of the genetic effects that habitat fragmentation may produce on marine populations.

Thierry Backeljau expressed his concerns about the application to marine environments of the concept of habitat fragmentation, developed originally for terrestrial systems. He argued that often people are tempted to study patchy populations as a model for habitat fragmentation and that, in this respect, it is important to study the population subdivision in a relatively short time span. Care should be taken when interpreting data, because habitat fragmentation is not the only phenomenon determining population subdivision.

Ferruccio Maltagliati replied by providing two examples. The first was about the disruption of one-dimensional gene flow in a hypothetical coastal species with low potential for dispersal, due to the presence of an even minor construction on the coast. The second was about fragmentation of seagrass meadows that may reasonably provoke population subdivision of organisms associated.

Tim Wyatt (Instituto de Investigaciones Marinas, Vigo, Spain) posed the question on the possibility of a temporal fragmentation, namely a sort of genetic discontinuity among cohorts of exploited stocks of fish. This could be provided by 1) the capture of the pool of high-ranking males that breed successfully and contribute to the progeny during spawning and 2) the exploitation of individuals whose presence is fundamental for the maintenance of the social structure because they are responsible of the control of the shoals for migration to spawning grounds.

Thierry Backeljau stressed the importance of studies that consider the impact of exploitation on breeder males. However, he expressed doubts on the analogies between the definitions of temporal and spatial fragmentation.

Andy Beaumont (School of Ocean Sciences, University of Bangor, Anglesey, UK) offered some comments on various aspects of the discussion. Firstly, he reported the importance of considering the effective population size in aquaculture, fisheries and conservation. Secondly, he argued that molecular biology advances have been mainly in methods that look closer at the DNA, rather than in increasing theoretical understanding of population genetics. This has produced a large temporal gap between the advent of a new genetic methods and their application for ecological or conservation purposes. Thirdly, he showed that micro-organisms and haploids are not taken into account when considerations on genetic diversity and conservation applications are made. Finally, he reminded us that human impacts may also homogenise populations, as happened with the European oyster.

Sebastian Holmes (Nederlands Instituut voor Onderzoek der Zee, Den Burg, Texel, Netherlands) argued that human-mediated transport of organisms exists and that it increases the levels of within population genetic diversity. In contrast, it decreases the degree of between population genetic diversity. He also stressed that the negative effects of inbreeding are overestimated. He separated the significance of within- and between-population components of genetic diversity and provided a terrestrial example where a recovery plan of a species has been successfully performed despite inbreeding. Furthermore, he argued that, from a metapopulation perspective, well connected (globally genetically homogeneous) populations that are locally genetically heterogeneous represent an evolutionary unstable strategy, in face of environmental variability. In contrast, poorly connected populations (globally genetically heterogeneous) but locally genetically homogeneous represent an evolutionary stable strategy.

Frédérique Viard (Station Biologique de Roscoff, Roscoff Cedex, France) agreed with this vision of inbreeding. She enlarged the concept stressing that, since inbreeding is common in nature, we should not blindly try to limit it in every situation, but we should assess the effects of increased relatedness on fitness. The analysis of bottlenecks or founder effects in natural populations may represent a useful tool to make reliable predictions for increased relatedness. Didier Aurelle (Centre d'Océanologie de Marseille, Station Marine d'Endoume, Marseille, France) joined the above ideas on inbreeding and introduced the importance of adaptive genetic variation at population or species level.

Keith Hiscock (Marine Life Information Network, Marine Biological Association of the UK, Citadel Hill, Plymouth, UK) stressed the importance to conservation of recovery, following decline, of a given species at a local scale, especially if the species is a key structural, key functional, or rare. He reported the different consequences of self-recruiting populations compared to populations recruiting from distant sources. He provided some examples of non-commercial invertebrates with different life history traits and population structure from British waters. He concluded by expressing the need of genetic investigation for those species in order to give insight on their conservation prospects.

Bella Galil (Israel Oceanographic and Limnological Research Ltd., Tel-Shikmona, Haifa, Israel) provided the example of a Mediterranean endemic reef-building gastropod with very low potential for dispersal. Given the distribution and life history traits of this species, she stressed the need of population genetic studies in order to shed light on its conservation status.

Another theme of the conference introduced by Ferruccio Maltagliati was environmental management and, in particular, genetic monitoring, meant as an array of methodologies to assess the response of populations to environmental stress. Genetic monitoring should be aimed at determining genetic loss and alterations of among-population genetic diversity. The effects of environmental stress, pollution, or habitat fragmentation on natural populations are examples of problems that could be addressed with genetic monitoring.

Lydia Ignaties (National Centre of Scientific Research "Demokritos", Aghia Paraskevi, Attiki, Greece) argued that biodiversity mostly relies on phenotypic characters of species. This leads to mistakes due to problems associated with species identification and taxonomy. She supported, however, the efforts to set up methodologies of genetic monitoring.

Piero Cossu (Dipartimento di Zoologia e Antropologia Biologica, University of Sassari, Sassari, Italy) joined the idea of the importance of genetic monitoring in environmental management. He proposed the Inter Simple Sequence Repeats (ISSRs) markers as a quick tool for the first step of the procedure of genetic monitoring.

Frédérique Viard (Station Biologique de Roscoff, Roscoff Cedex, France) expressed her ideas on the fact that the concepts of Management Unit (MU) and Evolutionary Significant Unit (ESU) should be enlarged by including kinship, neighbouring size and mating system.

Adriana Zingone (Stazione Zoologica 'Anton Dohrn', Naples, Italy) with her interesting experience on phytoplankton species reported that coastal waters should assume greater conservation and protection relevance. In fact, coastal waters can be considered a sort of source of genetic diversity for phytoplankton organisms and they constitute, therefore, evolutionary grounds for many species. Human impacts on coastal waters may provoke the disruption of the phytoplankton 'genetic diversity maximiser' with a consequent drop of genetic diversity.

Recognizing the "units" to assess, chart and measure biodiversity, is pivotal for management and conservation purposes. Yet, the discussions on this issue clearly showed that there still exist a number of controversial points, leading to lively exchanges between several participants.

Thierry Backeljau emphasised the importance of going beyond the species approach in measuring biodiversity by considering infraspecific genetic variation for it is this infraspecific variation that provides evolutionary potential to organisms. In this context he referred to the practical value of concepts such as "Evolutionarily Significant Units" (ESUs) and "Management Units" (MUs) to chart infraspecific genetic variation in relation to evolutionary history. Yet he also emphasized the importance of quantitative genetic variation , in addition to molecular data and more "classical" characters. Nevertheless currently quantitative genetic analyses are seldom included in biodiversity research or in the application of ESUs and MUs (neither of these latter two concepts accommodate for quantitative genetics).

Part of these statements provoked reactions from Ferdinando Boero (Dipartimento di Scienze e Tecnologie Biologiche ed Ambientali, Università di Lecce, Lecce, Italy), Ward Appeltans (Flanders Marine Data and Information Centre, Flanders Marine Institute, Oostende, Belgium), Marco Casu (Dipartimento di Zoologia e Antropologia Biologica, University of Sassari, Sassari, Italy), Godtfred Høpner Petersen (Zoological Museum, University of Copenhagen, Copenhagen, Denmark), Yves Samyn (Laboratory of Ecology and Systematics, Vrije Universiteit Brussel, Brussel, Belgium), and Stefano Mariani (Department of Biological Sciences: Molecular Ecology and Fisheries Genetics Laboratory, University of Hull, Hull, UK) who emphasized the central role of species as unit to measure and chart biodiversity. Ward Appeltans, for example, stressed the need for species lists as a tool to recognize biodiversity hot spots and endemisms. Although this point was not contested, Andy Beaumont and Thierry Backeljau remarked that (1) applying species concepts to several organisms remains problematic (e.g. in asexuals) and (2) that "species lists" often are very biased in the part of biodiversity they cover. Usually they include only "larger" metazoans and plants, and often they do not consider prokaryote, protist and meiofaunal diversity, even if this may be an important component of the overall biodiversity in a given ecosystem. Moreover, these groups of organisms are poor in phenotypic characters by which the organisms can be identified. Hence their recognition and diversity assessment is by necessity largely based on molecular genetic techniques.

This latter point, together with the earlier statement of Thierry Backeljau concerning the importance of infraspecific genetic variation, opened another discussion in which concerns were expressed about the perceived controversy between "phenotypically based" and "genetically based" approaches to biodiversity (e.g. Ferdinando Boero, Ward Appeltans). Everybody agreed that both approaches need each other and have to be integrated. This requires of course a good communication between researchers that may have very different backgrounds and interests (e.g. Ferruccio Maltagliati, Yves Samyn, Ferdinando Boero, Thierry Backeljau). Some participants felt that on this point there still is a long way to go (e.g. Ferdinando Boero), while others perceived this a lesser problem given that they precisely combined molecular genetic and phenotypic approaches (e.g. Stefano Mariani, Thierry Backeljau)

In a similar way, the discussion on species concepts arrived at the general conclusion, that there is no such thing as a generally applicable species concept, even if everybody recognized the relevance of delimiting species and this latter of course requires some interpretative framework under the form of some sort of species concept. Nevertheless, Godtfred Høpner Petersen did not endorse this latter point and stated that "species are" and "exist without definition as components in different food webs and energy flows". Most other participants, however, seemed to attach importance on the consequences and scientific, as well as practical, implications of applying one or another species concept. Ferruccio Maltagliati remarked in this context that the discussion about species concepts is an old one and that each time new analytical tools are introduced to taxonomic research, there is some sort of initial euphoria related to the hope that the newer techniques will solve the old, long-standing problems. This applied to protein electrophoresis at the end of the 1960s and to PCR combined with DNA sequencing at the end of the 1980s. In both these cases it took several years before it became clear that no technique can be considered as Pandora's box, i.e. each technique has its shortcommings and limitations with respect to measuring genetic variation and phylogenetic relationships. Hence no technique can solve all (taxonomic and phylogenetic) questions and problems (e.g. Andy Beaumont). As such molecular techniques have added to the species concepts problem, but have certainly not solved it (perhaps they even have aggravated it).

Yet this resulted in another concern of Ferdinando Boero and Peter Mill (The School of Biology, University of Leeds, UK), viz. that currently students, university curricula and funding agencies show little interest in "classical" (i.e. phenotypically based) taxonomic and ecological work. As a consequence, a lot of taxonomic expertise may be lost. However, Stefano Mariani and Thierry Backeljau were more optimistic about this, for they claimed that "classical" approaches are subject to a revival, both among students and funding agencies.

Perhaps the major take-home message from the discussion on taxonomy and species concepts, is that one might easily devote an entire conference to these issues, probably without arriving at a general consensus. Yet this seems inherent to the very nature of biodiversity itself: it seems over-simplistic to try to look at biodiversity from one single point of view and hence the synergy between different approaches and conceptual frameworks is probably the best way of dealing with the measurement of biodiversity.

Having read all of the contributions to the conference it is clear that the community of scientists reached by the medium of electronic communication are fairly complacent about the potential effects of aquaculture on marine fish and shellfish. There were few direct comments on the subject of aquaculture.

Is there a need to worry about the potential effects?

As mentioned in the opening document aquaculture is the only sector of food production that is continuing to grow. In Europe and many other developed countries this is because fish is displacing other items from the diet, mainly on the basis of its healthy image. At present the world's aquaculture production is dominated by the production of freshwater carp species in Asia. This form of aquaculture is generally seen as environmentally neutral or possible beneficial. Much of this production is actually from species that are non-indigenous to the countries in which they are grown.

However, the new growth in the aquaculture sector will be in the development of marine fish and shellfish because of their economic value and greater availability of suitable farm sites. The rapid and massive expansion of farmed marine shrimps and the devastation associated with the loss of mangrove habit being a prime example of unsustainable aquaculture development.

It has been the problems of reproducing the broodstock and rearing the larval stages that has hindered the growth of the industry. However, the techniques applied to rearing seabream and seabass are now proving successful in other marine species such as cod and halibut so reducing the time taken to close the life-cycle of new species.

Once successful hatchery systems have been developed the actual number of fish in the industry may be many times that of the natural populations of a given species (eg Atlantic salmon production now nearly 1.2 million tonnes world-wide approximately a European wild catch <5000 tonnes). Widespread escapes under these conditions would eventually result in a homogenisation of allele frequencies over quite large areas of the species range. The seabass and sea bream industry standing crop (>100,000 tonnes) must be close to that of the total natural population size of both these species.

The inability of politicians to manage wild fish stock will mean that aquaculture will continue to produce more and more of the fish we eat. Farmed cod is already preferred by the premium hotel and catering trade because of its freshness, fillet quality and taste over wild fish.

To date few restriction have been imposed on the provenance of the fish being used in any form of aquaculture, recent legislation in the USA restricting the salmon industry in Maine to only use indigenous salmon strains is one of the first. This will put the local salmon farming industry at a distinct disadvantage, as these strains will perform much worse that the selectively improved strains used by farmers elsewhere in the world. Is this a long-term solution? Hatchery strains will change as they become domesticated unless managed as ranched stock.

The comments by Ferdinando Boero questioned the sustainability of the industry based on present feeding regimes. He likened the present process to us growing grass to feed sheep to feed these to wolves so we could feed lions. Fish are actually very efficient converters of food into protein compared to mammals and birds as they have a small metabolic cost. Substitution of fish meal and oil by vegetable protein and fat is feasible but marine fish cannot elongate or desaturate the fats into the omega 3's that we are being encouraged to eat because of their health giving properties. Most feed companies are trying to develop a standard that will result in 1kg of fish meal producing 1kg of farmed fish by controlled substitution. Fish meal is still relatively cheap and the cost of fish food is relatively expensive so the large feed companies will be happy to supply this sector over the price sensitive chicken and pig sectors. If we continue to loose our large carnivorous species what will eat the small prey species such as sandeels, capelin, anchoveta etc it may as well be turned into fishmeal if it is not polluted by PCBs.

The comments on the Conover and Munch work would not be surprising to people involved in selective improvement of fish. Properly controlled breeding experiments are showing good heritabilities for many traits important to aquaculture such as growth, body conformation and disease resistance. The fact that we are still dealing with essentially wild animals with relatively high levels of genetic variation and can impose high selection differentials, because of the high fecundity of fish, will ensure that these gains are likely to be maintained for many generations. Estimated improvement in growth performance of Norwegian salmon over 9 generations of selection is > 100%.

From the bullet points it can be read that aquaculture will continue to grow and the number of species being farmed will continue to increase. Improvements in husbandry, nutrition and genetics will result in improvements in the efficiency and sustainability of the industry. The potential impacts of aquaculture will be different to that of fisheries. Aquaculture could be used to mitigate the over exploitation caused by fisheries by releasing young fish to enhance natural stocks. This is already a widespread practice for salmonids and many freshwater species and is now being used in marine fisheries for lobster and Penaeid prawns. The availability of large number of animals for experimental purposes, particularly if they are from pedigreed stocks, is an important scientific asset to undertake genetic analysis of important fitness traits that have often been neglected in favour of a more molecular approach.

Aquaculture will result in escapees from most production units. This may result in the release of species that are non-indigenous and possibly invasive, such as the expansion of salmonids in South America. The use of indigenous species will mean that escapes will reduce possible differentiation between populations as gene flow increases between domesticated and wild populations. Avoidance of both these scenarios is possible by the use of single sex and/or sterile stocks that are relatively easy to produce, but are presently not acceptable to the industry.

Fish farms may also alter the marine environment at least locally because of the infrastructure, nutrients released and interactions with the same or other species. Some of the potential problems such as build-up of detritus are now actively managed by fallowing, predation by large carnivores (seals sealions) is being controlled by better net designs. But this is new industry and we have little understanding of possible longterm effects that might appear. However the aquaculture industry is very sensitive to public opinion and, unlike fisheries, the larger more responsible companies will respond quickly to issues as they appear. If we continue to deplete our wild fish stocks and we want to continue to eat ever more fish then we will have to find the answers to the problems as they appear.

I recommend three particular papers that I believe are balanced assessment of the potential and problems of this new industry.

The marine realm is a theatre in which some very large scale selection experiments are being played out, with consequences that are ill-understood and potentially serious for the future health of our marine resources. The actors in this play are: (1) the fishery managers who set the rules by which fish are selected, (2) the fishers who apply the selective mortality, and (3) the fish stocks that undergo genetic change due to selective fishing.

The basis for this argument is as follows: all fishing is selective. Most obviously the market attaches different values to fish of different sizes, and fishers try to catch sizes that give the greatest return. In addition fishery managers try to regulate fishing through technical measures such as net mesh sizes. The trait most evidently under selection is size-at-age. But there are many others, including size and age at sexual maturation, potentially important for productivity of fisheries.

The strength of selection is large. Most fisheries are fully exploited or over-exploited (FAO, 1998); fishing mortality rates are often two or three times greater than all other sources of mortality put together. The strength of selection on body size is large enough to be detectable within single year classes of North Atlantic cod, as the fish grow and selective mortality takes place (Sinclair et al., 2002).

There is genetic variation in the traits under selection. Over the last 20 years, the aquaculture industry has shown that it is quite feasible to bring about genetic change in growth and maturation of fish in captivity; this can only happen if these traits contain genetic variation. The heritabilities are in the range expected of traits closely related to fitness (Mousseau & Roff, 1987), even when estimated directly in the wild (Jónasson, Gjerde & Gjedrem, 1997).

Measurable changes in the traits are observable at least on a time scale of decades. Large phenotypic changes are taking place in major commercial fish stocks. These changes can, of course, be due to factors other than genetic change driven by selective exploitation. Nonetheless, the changes are often in a direction consistent with evolution, and remain even when proximal factors such as food supply and temperature are factored out (e.g. Rijnsdorp, 1993; Grift et al., 2003; Heino, Dieckmann & Gødo, 2003).

The traits that change under fishery-induced selection are important for the sustained productivity of exploited fish stocks. Most startling are the results of a laboratory selection experiment on the Atlantic silverside (Conover & Munch, 2000). Starting from the same stock, they culled either small individuals or large individuals from replicate populations; within four generations, the yield from the former was nearly twice that from the latter.

E-conference debate

Responses to the argument above mostly addressed the issue of how to manage fisheries, given that they evolve under exploitation. One view was that a substantial part of the marine realm should be designated as marine reserves (Mariani 9 October; 10 October), leaving natural processes to their own devices. There is much to recommend this, although the effectiveness of reserves clearly depends on the amount of gene flow across their boundaries and the strength of selection generated by fishing, relative to natural selection. [Theory suggests that the strength of selection generated by fishing can be at least an order of magnitude greater than natural selection in the reverse direction (Law, 2000, fig 3b).] Also, the merits of reserves have to be balanced against economic interests of communities dependent on fishing for their livelihoods.

Another issue was the extent to which the experimental results of Conover and Munch (2002) can be used make inferences about evolutionary management of stocks in the wild (Mariani 9 October; Kraak 10 October; 15 October). The life history of the species used by Conover and Munch was somewhat different to that of many major fish stocks; the work demonstrates the existence, but not necessarily the quantitative path of evolution under exploitation. It would be a great help to put in place experiments of a larger scale on commercially important fish species (Kraak 10 October).

Is it better to harvest smaller or larger fish for evolutionary management (Kraak 9 October; Mariani 9 October)? I suggested that there is some merit in maintaining a stock of large fish having relatively low fishing mortality rates on the basis that this helps to maintain the spawning stock biomass in the short term, and may generate selection for faster growth in the long term. However this argument runs counter to conventional view that fishing mortality should be concentrated on larger fish that have already achieved much of their potential growth. Clearly reducing fishing mortality on large fish will not do much good if it simply results in greater fishing mortality on smaller fish (Kraak 9 October; Law 11 October). There is a real need for research into this issue (Kraak 9 October; 15 October), guided by theory available from life-history evolution already in the literature (Law 11 October).

Another point raised in discussion was the erosion of genetic variability through small effective population size (Charrier 9 October). Whether there is general evidence that fishing causes small enough population sizes is open to question (Law 11 October). A major loss of variability requires dramatic genetic bottlenecks (Maltagliati 14 October), and populations driven to such small sizes by fishing may be on their way to extinction in any event. Complete destruction of the gene pool in this way is obviously of special concern.

Conclusion

The precautionary principle places a responsibility on us to leave our marine resources in a state that can be utilised as fully by our descendants as by ourselves. There is little systematic thought being given to these matters of selection and genetic change caused by fishing at the present time. If we continue on our present path, our descendants are not going thank us. Arguably we need a new - Darwinian - fisheries science.

Co-chairs introductions to Topic 3 focussed on two major aspects related to human alteration of gene flow in marine environment.

The first issue is biological invasions. Species-range expansions - often through human-mediated transport - have been observed at an increasing rate since the end of the 19th century. Biological invasions are one of the major threats for marine biodiversity and ecosystem stability. Several authors have underlined the potential usefulness, but lack, of molecular data for the study of biological invasions. Genetic studies could provide the opportunity to address major issues including the origin of newly founded populations, the rates of dispersal, and pathways through which dispersal occurs. Further contribution are to depict and understand the rates of change in the genetics of non-native species and associated life-history traits, the importance of natural selection in shaping the genetic architecture of genetically impoverished populations, and the degree of hybridisation between native and non-native species.

The second point addressed in the introduction dealt with human induced changes in the metapopulation structure of marine species caused by alteration of coastal habitats. Marine systems are often assumed to be continuous, and marine species to be able to disperse over large geographic ranges, yet the effective long-distance dispersal ability of marine species has rarely been quantified by direct measurements. Many marine species contradict this assumption, and show remarkable levels of genetic distinction over a range of geographic scales. Human activities can modify natural patterns of coastal landscapes, thus altering the level of isolation and connectivity among populations. They act both by the creation of corridors that allow/increase species dispersal, and by changing the degree of habitat fragmentation, thus affecting evolutionary processes. How relevant are these issues in the marine environment and how could population genetic studies provide insights into gene flow models that are relevant to marine species?

Several lines of discussions were followed during this topic and the two chairmen are grateful to all the contributors for their fruitful and stimulating discussions. From the 19 messages received during the three-days devoted to Topic 3, several important issues were addressed.

Firstly, the effects of human activities in altering genetic characteristics and gene flow in the marine environment are still to be quantified. In that context, the usefulness of population genetics is certain. However the time-scale needed to differentiate between 'natural' processes and anthropogenic effects is a critical point and the applicability of population genetics for short-term management purposes is debated. There is still work to be done to formulate hypotheses that includes an a priori expectation as to the relevant scale to be considered and protocols useful for management purposes (see contributions by M. Abbiati, J.-P. Féral, Ferruccio Maltagliati). For instance, we cannot expect to see the same evolutionary and genetic patterns when considering climate change when investigating the sudden human-mediated introduction of a new species in a given habitat. Ferruccio Maltagliati also opened an interesting discussion concerning both the rate of evolutionary processes and the relevance of theoretical models (e.g. metapopulation) used primarily in the terrestrial environment (see also contributions by Frédérique Viard, J. P. Féral, M Abbiati).

Several contributors highlighted the need for extending our basic knowledge in systematics, taxonomy, phylogeography and species recognition (see the discussion line initiated by Keith Hiscock and answers from John Bolton, Sandra Duran and Frédérique Viard). This appears as a major issue concerning invasive species in particular when dealing with the concept of cryptogenic species (see Carlton 1996) and has implications for the establishment of species lists (e.g. list of threatened species) and management policy. For that purpose, as underlined several times not only in this topic but also in topic 1 and 2 of this conference, molecular and morphological analyses should both be carried out.

A third line of discussion dealt with the effects of climate changes (see contributions by Pierre Mathy, Boris Winterhalter, Keith Hiscock, Ferdinando Boero). Several programs (eg MarClim) and international councils or organizations (eg CIESM) have reported evidences of climate change on various biotas and some contributors provided specific examples. The genetic consequences of climate change at the species-level are however not straightforward (but see, for instance expectations about founder effects due to colonization of new habitats).

Many questions have been examined and many others are still to be discussed (see the questions addressed in the introduction by the chairmen). Obviously, marine communities and species are threatened by human activities. Also without a doubt, as with species in other environments, selective pressures (due to pollution), disruption of local populations, decreases in demographic size and geographical range of populations and the translocation of populations have implications on the genetic diversity and structure at the species level. All this is likely to affect species evolution. However, again without a doubt, the scientific community needs to define and formulate more specific and practical tools for management purposes and, at the same time, has to continue to extend its basic knowledge about marine species and communities.

I hope that you have all enjoyed and learned something from the MARBENA electronic conference on 'Genetic biodiversity in marine ecosystems: measurement, understanding, management'. Although I feel something of a fraud overseeing a conference on genetics (I am a simple marine ecologist who likes whole organisms and understanding how they are distributed according to environmental conditions), I hope that we have managed to structure the conference in an informative and thought-provoking way.

I have learned a terrific amount particularly about the power that genetic studies have now and will have in answering questions about how humans are affecting the marine environment. The discussion about morphospecies and what we can learn about underlying biodiversity from molecular genetics was of fundamental importance in how we develop taxonomy (including in support of conservation) in the future. Although I agree with some later statements that morphological characteristics will continue to be the most practical tool for the field ecologist. I hope that the conference has given you some ideas about how studies of the genetics of marine organisms can help to identify those human activities that may most damage the marine environment. Those mentioned activities ranged from introduction of non-native species, through impacts of climate change and coastal construction to the effects of aquaculture and fisheries. 'Connectivity' and the role of human activities in breaking-down barriers has been a theme throughout the conference.

The concerns that pre-occupy agriculture on land - especially genetic manipulation/modification- do not seem to over-occupy marine biologists. However, significant worries about the impacts of fisheries have been raised - contributing to the now accepted view that fisheries are damagingly out of control and over-fishing may have permanent or very long-term adverse consequences. Some more discussion on genetic effects of fisheries and of aquaculture than our conference has given might be appropriate - commercial interests operate much more quickly than careful 'academic' research and before you know it, there is a problem.

Finally, my thanks to Federica Pannacciulli and Isabel Sousa Pinto for helping to determine topics that would inform the physical conference in Florence on "Genetic Biodiversity in Natural and Agricultural Ecosystems" being organised for 20-24 November 2003. Thank-you to the co-chairs for the three topics for providing such good outlines and for asking questions that provoked you into replying. Thank-you Ward Appeltans at the Flanders Marine Data and Information Centre for keeping the show on the road so effectively. And, especially, thanks to all who contributed their thoughts.