(Eric Collin)

CEMAGREF - Groupement de Nogent-sur-Vernisson - France
 

Why should gene conservation of elms be treated differently from that of other common Noble Hardwood species? The sole answer that readily comes to mind is: ‘because elms are endangered species, due to Dutch elm disease’. As a matter of fact, the two Dutch elm disease (DED) pandemics that spread across Europe this century have caused dramatic mortality in elm populations, and the disease still represents a great threat for each individual elm tree. However, the statement that ‘elms are endangered species’ needs to be scrutinized and the situation of each elm species must be considered separately. Factors of importance in elm gene conservation must be discussed before practical recommendations for the implementation of elm gene conservation can be made, in the light of the principles and methods defined for the other common Noble Hardwoods.

This is a two-fold question: how many elm species are there in Europe, and are they in danger of extinction or at least under threat of severe genetic depauperation ?

Because the taxonomy of elms has always been a bone of contention among botanists, it is difficult to give a simple answer to the first question. Some botanists consider that there are only three elm species in Europe (Richens, 1983), and that major variants in these species must be given the rank of varieties or sub-species. Others (Melville, 1940 and 1957) advocate the recognition of species when regional forms display a very distinctive habit (e.g. Ulmus plotii Druce in the English Midlands) or conspicuous characters such as leaf hairiness (e.g. U. canescens Melville in the Eastern Mediterranean region) or seed hairiness (e.g. U. elliptica Koch in the Caucasus). Little information is available about U. celtidea, an endemic Russian species differing from the European white elm (U. laevis Pall.) in tree size and form of the samaras (Yarmolenko, 1936).

The second question can be answered more easily : elms are not endangered at specific level, except if one takes the ‘species’ concept in an extremely narrow sense. For instance, because of its status as an endemic, U. plotii is listed in the UK Biodiversity action plans as a species of high conservation importance. However, such a narrow ‘species’ concept may prove misleading, as demonstrated by recent molecular work (Coleman et al., in press) showing that U. plotii is in fact composed of a single widespread clone mixed with a number of morphologically similar but genetically unrelated entities. Conversely, some rare forms and marginal populations of European elms are indeed in serious danger because of DED and/or the destruction of their habitats. More generally, there are risks that repeated tree losses and the reduction in the number of genets effectively contributing offspring might result in genetic depauperation in the long term. Besides, as stated by Millar and Libby (1991), perhaps the greatest challenge to conservation biologists is not only to "rescue species from the brink of extinction" but also "prevent currently healthy taxa from ever becoming endangered".

From a practical point of view, we suggest that the European elms be regarded as three ‘large species’ (see section below) with different biological and ecological features, and thus requiring different gene conservation strategies. The present knowledge on the taxonomic and geographic partitioning of genetic variation within each of these large species is fragmentary and sometimes controversial. However, it provides a basis for the implementation of a pragmatic combination of in situ and ex situ conservation measures as well as indications for further research. For instance, regardless of names and putative taxonomic ranks (‘U. glabra Huds. var. trautvetteri Johansson’ as opposed to ‘U. elliptica Koch’), one should investigate the genetic diversity in the distinctive Caucasian form of Wych elm (U. glabra) and its genetic distance from the type devoid of hair on the seed. Similarly, even if one may doubt that the Mediterranean forms of Field elm (U. minor Mill.) with densely hairy leaves really compose a genetic entity, it is certainly worth while to preserve in ex situ collections clones representative of the small populations of U. canescens that still can be found in Malta and other places.

We do not think that the principles of elm gene conservation differ fundamentally from the general guidelines proposed for the other common Noble Hardwoods (Jensen et al., 1999). In all cases, rarity and endangerment do not occur at the species level but rather at intraspecific level. Just as for other common Noble Hardwoods, our goal is to ensure the conservation of the evolutionary potential of the elm species rather than strictly preserve the present state of their gene pools. The only difference is not in terms of goals but in terms of difficulties encountered. From this point of view, Dutch elm disease poses probably the most critical problem, but certainly not the only one. As will be discussed below, habitat fragmentation is also a major issue in the conservation of marginal populations of the European white elm (U. laevis). Ancient transfers by humans and a good ability for vegetative propagation have combined their effects to render sampling for conservation particularly difficult in countryside Field elms (U. minor Mill.). So, the case study of elms should certainly not be reduced to its pathological component but rather be viewed as a condensed example of the many kinds of difficulties that can be met when undertaking the genetic conservation of forest tree species.

Ulmus glabra.

Wych elm (U. glabra Huds. = U. montana With. = U. scabra Mill.) is native to most European countries. It thrives at northern latitudes, but also occurs in the montane forests of southern Europe. As latitudinal variation is strong in U. glabra, Uotila (1997) suggests to recognize two subspecies, a northern one and a southern one. Apart from the case of U. elliptica already discussed (see above), interesting taxonomic variants with a corky rhytidom were described in Rumania (Borlea, 1995) and a distinctive form of U. glabra with long leaf stalks occurs in northern Spain (Richens, 1983). In Sweden and Norway, rare relict populations of U. glabra can be found not very far from the Arctic Circle (L Ackzell, pers. com.).

Ulmus glabra is very susceptible to DED, but remains unaffected at the northernmost latitudes where no efficient insect vector is available. The threat to its genetic resources in the central part of its natural range is still debated among specialists. In some areas, it is not considered endangered because it regenerates easily from seed. In other regions, foresters report that, because U. glabra does not set seed regularly, the trees that die of DED are often not replaced by younger ones. Moreover, whenever seedlings appear, they are likely to be destroyed by deer and other game. In Germany, foresters have started implementing practical preservation measures such as fencing seedling patches. If natural regeneration is missing, then they plant material obtained from seed lots harvested locally. The situation of small marginal populations requires special care because they could be devastated by DED in a very short time. This seems to be the case in Sicily (L Mittempergher, pers. com.).

The main ecological characteristic of U. glabra to consider in a sampling strategy for conservation purposes is its ability to form fairly large sexually reproducing populations. Because it is a wind-pollinated species, gene flow and random mating are expected to be important, so the risks of inadequate sampling are probably limited due to the large ratio of within- to among-population variation. However, at the margin of its distribution, U. glabra is scattered, and its conservation should be targeted with a greater precision.

Ulmus laevis.

The European white elm (U. laevis Pall. = U. effusa Willd. = U. pedunculata Foug.) belongs to section Blepharocarpus, which also comprises its American relative, U. americana (American elm). Ulmus laevis probably does not hybridize with U. glabra and U. minor (Mittempergher and La Porta, 1991). It has a definite east European distribution, but it reaches eastern and central France. There are even a few relict populations in southern France (Timbal and Collin, in press.). Ulmus laevis is a component of riparian forests along large rivers such as the Rhine, the Elbe, and the Danube. It is one of the rare European tree species that thrive in damp soils flooded at some periods of the year.

U. laevis is susceptible to the fungal agent of DED (Ophiostoma novo-ulmi), but, at least in western Europe, its populations are generally not severely damaged by the epidemics. Experiments with captive elm bark beetles (Scolytus scolytus and S. multistriatus) allowed to feed on different elm species have shown that U. laevis is far less attractive for the elm bark beetles than U. minor (Sacchetti et al., 1990 ; Webber, 1999 ; D Piou, pers. com.). As long as the insect vectors do not change their feeding preferences, the genetic resources of U. laevis are not really endangered by DED. In fact, U. laevis is certainly more seriously threatened by the destruction of the riparian forests where it can be found. Major changes in the landscape are likely to happen along the banks of large rivers, especially when the land can be drained and reclaimed for agriculture. The changes in land ownership in many east European countries with economies in transition have also threatened riparian woodlands. Thus, conservation of the genetic resources of U. laevis should concentrate on marginal populations (e.g., in southern France) and on remarkable floodplain communities in danger of deforestation. It could be undertaken either in situ, when some kind of legal protection of habitat is available, or ex situ as a complementary measure.

Ulmus laevis is often found in fragmented populations of limited size that may be subject to serious risks of genetic drift. Studies conducted in Finland with isoenzyme markers (Mattila and Vakkari, 1997) have suggested that random genetic drift may have caused substantial differentiation among the small populations of U. laevis at the northern fringe of its range. Hence, sampling for adaptedness should be carried out very carefully and focus on populations with a large genetic diversity. On the other hand, populations that are likely to have encountered high levels of genetic drift should also be targeted in order to maximise the capture of allelic diversity.

Ulmus minor.

Field elms have been given so many Latin names (Richens, 1976) that some preliminary explanations need to be given here. In theory, the binomial U. campestris L. should be rejected because it is an ambiguous name since Linné did not make distinctions between Wych elm and the Field elms (Melville, 1938) ; however, it is still popular among botanists and foresters. The binomial U. carpinifolia Gled. is frequently used to designate the diverse forms of Field elms that are common on the European mainland ; it does not include the English Elm (U. procera Salisb.). Following Richens (1968), we favor the enlarged circumscription of the binomial U. minor Mill. sensu latissimo, which covers the Field elm complex in its totality ; in this treatment, forms with distinctive branching habits and leaf shapes are given a varietal rank. The many other Latin names that were given to Field elms, such as U. nitens Moensch. and U. foliacea Gilibert, are liable to bring more confusion than clarity and should not be used any more.

Ulmus minor differs from U. glabra and U. laevis in many respects. First, U. minor is adapted to warmer climates. Its distribution is clearly south European, but also occurs in Algeria, Turkey and northern Iran. It is probably not native in England and northern countries, and its status on the Baltic islands is problematic (Richens, 1983). The second specificity of U. minor is due to the role of humans (Heybroek, 1990, 1993b). Whereas U. glabra and U. laevis are mainly found in natural woodlands, U. minor has been widely propagated and planted during the last two thousand years, and probably even earlier. For instance, according to Richens (1983), the regional variety known as Cornish elm (var. cornubiensis, syn. U. stricta) was introduced from western Brittany into Cornwall before the ninth century. Similarly, the English elm (var. vulgaris, syn. U. procera) may have been introduced from the Iberian peninsula into Britain in the Later Bronze Age. The third major difference between U. minor and the two other European elm species is its large variability and taxonomic complexity. An illustration of the biosystematic and nomenclatural difficulties that arise in the U. minor complex is given by the conflicting treatments proposed by eminent elm specialists for the different kinds of U. minor found in Britain (Armstrong and Sell, 1996). Whereas Melville (1975, 1978) recognized them as five different species, several varieties and many complex hybrids, Richens (1968, 1980) pooled them into a single large species with four varieties. In the microspecies approach adopted by Armstrong (1992) they are described under about 40 binomials, most of them new ones. The complexity in the present day distribution of the British elms is partly due to their ancient introduction and cultivation by humans. However, natural variability in U. minor is particularly high and poorly known. Richens (1980) mentions a longitudinal pattern of variation which could be explained by northwards migrations from "a chain of distinct Mediterranean refugia during the last Glaciations". Hopefully, molecular markers should help clarify this taxonomic chaos (Hollingsworth et al., 1999) and provide evidence on ice refugia and possible re-colonization routes followed by the elms.

Ulmus minor hybridizes easily in nature with other elm species and varieties (apart from U. laevis). Its hybrids with U. glabra compose a very large and diverse group of intermediate forms commonly referred to as U. x hollandica, - i.e. the Dutch hybrid elms. It seems that, in Spain and Italy, U. minor is now being introgressed by the introduced Asian species U. pumila. This very recent introgression process may be regarded as a potential asset as far as tolerance to DED is concerned, but people fear that it could also be detrimental to the quality of wood.

Another feature of major importance for conservation biologists is the excellent ability of U. minor to produce root suckers and to re-sprout from the stump. The first consequence of this ability is that U. minor can colonize new sites by seed and then rapidly reproduce by root suckers, eventually forming thickets with little or no genetic variation. Another consequence is that vigorous suckering and re-sprouting occur very frequently in countryside elm hedges after they have been devastated by DED. Thus, in spite of the heavy losses due to the disease, the genetic resources of the common varieties of Field elms are probably not as threatened as thought at the peak of the epidemics when all mature trees were dying so quickly.

Because of the taxonomic confusion in the U. minor complex, the occurrence of clonal-patches in natural or sub-spontaneous stands, and the gene transfers caused by its ancient cultivation, sampling for genetic diversity in U. minor is particularly difficult. Different measures should be applied in the case of riparian natural stands versus the more common situation of elm landscapes largely influenced by human activity.

A detailed review of the results of research work carried out on elm biology, biosystematics and genetics cannot be presented here, but fields of major activity will be briefly highlighted.

Variation in macroscopic characters.

In the period 1940-1980, extensive studies of leaf variation in U. minor were made in several western European countries by two English botanists who had divergent views on the taxonomy of elms. Unlike Melville who employed a traditional approach of classification, Richens in collaboration with Jeffers developed a numeric method based on computer aided multivariate analysis of measurements of leaves. Richens and Jeffers published their results for England (1978 ; Jeffers and Richens, 1970), northern France (1975), Wales (1985) and northern Spain (1986). Richens was working on the interpretation of the results for former Yugoslavia when he died. In Spain, Ipinza (1989) employed the same method as Richens and Jeffers, whereas in Britain, Armstrong (1992) adopted a microspecies approach of British elms. In a recent paper (1999), Jeffers recalls Richens’ wish that their data (either the complete set obtained from over 2000 trees, or the ‘reference collection’ composed with a subset from 66 trees representing the full range of variation) be re-analyzed by other scientists with today’s computers and programs.

Apart from foliar traits, botanists have paid attention to a large array of characters such as: flower, fruit (e.g. size of samara, position of seed, form of seed notch, pubescence), bark and cork. However, no extensive collections of data were made at an international scale for such characters. Branching habit and aspect of rhitydom have been studied and illustrated for the most distinctive British varieties (Jobling and Mitchell, 1974).

Associated species.

Richens (1963, 1983) considered that monophage species associated with elms, such as Stigmella sp. (Lepidoptera), could provide information on post Glacial-migration. Similarly, he regarded the gall mites Eriophyes sp. as valuable indicators of the origins of the diverse local forms of elms transported by man in historic or prehistoric times.

Variation in microscopic characters.

The observation of chromosomes shows that the American elm (U. americana) is tetraploid whereas the other elm species are diploid. This last assumption was challenged by Machon et al. (1995) who concluded from isoenzymes profiles that the three European elm species must be segmental tetraploids ; however, other explanations were recently proposed for such profiles (Hollingsworth et al., 1999. ; M.A. Cogolludo, pers. com.).

Pollen lots of four European ‘species’ of elms (U. procera included) were discriminated under electronic microscopy by Stockmarr (1970, 1974), using the different frequencies in number of pores per grain observed in each lot ; U. procera was the less difficult to discriminate from the other species, even from U. minor. In the same study, fossil pollen from Lithuania indicated that U. laevis was the first elm species to re-colonize this part of Europe after the glaciations. However, according to Richens (1976), such investigations were based on too small samples to allow secure conclusions on the nature of palynological variation within the field elm generally.

Variation in elms, and particularly European elms, has been analyzed using flavonoids (Bate-Smith and Richens, 1973 ; Heimler et al., 1990 and 1993), isozymes (Machon et al., 1995, 1997), RAPDs (Coleman 1998, Coleman et al., in press), chloroplast and nuclear encoded ribosomal DNA restriction site analysis (Wiegrefe, 1992 ; Wiegrefe et al., 1993). A brief review of some of these studies can be found in Hollingsworth et al. (1999). More research is being carried out in the RESGEN 78 project funded by the European Union (see infra) as well in the frame of the Spanish national program. A Ph.D. study focussed on marginal populations of U. laevis is starting in Sweden and should involve the development of microsatellite markers for elms. A comprehensive review of the results obtained with biochemical and molecular markers needs to be made in the next few years.

Biologic traits and adaptative markers.

Botanists (Richens, 1983) have reported phenologic differences in bud burst and leaf fall between the early flushing English elm and other varieties of U. minor (such as the late flushing var. cornubiensis) in Britain. Breeders (Heybroek 1993a and 1999 ; Mittempergher and La Porta, 1991) have collected observations on the flowering and crossing ability of a large and diverse array of elm clones, and systematically investigated their susceptibility to DED in artificial inoculation trials. A few provenance tests were planted for Asian elm species of interest in breeding, but none for European elm species. Phenologic notations will be carried out in clonal banks and clonal tests in the frame of the EU RESGEN 78 project, and the Swedish Ph.D research mentioned above should include a greenhouse experiment allowing measurements of open progenies of U. laevis obtained from Sweden, southern France and a place in central Europe.

Pollen and seed dispersal, as well as self-sterility, are of major importance in the definition of gene conservation strategies for the species concerned. According to Walter (1931) and Daumann (1975), the three European elm species are normally wind pollinated but honeybees contribute to pollen dispersal and their role might be not negligible in the absence of wind. Elm seed is attached to a winged samara and thus easily dispersed by the wind. The samara is of medium size in the case of U. minor, larger in U. glabra and smaller in U. laevis. Heybroek (1993) reports that "elms are mostly to highly self sterile" and that "generally, random parents give better germinating seeds and more seedlings per pollination bag in crosses than in selfings, on average 23 times as many. Relations vary strongly from year to year however, and certain clones are relatively self-fertile". Heybroek does not give separate figures for the European species and the Asian species, and his data probably include clones with a complex parentage involving many species.

Richens (1983 ; Richens and Jeffers, 1978) made use of archeological findings and places names to trace up the introduction of elms. For instance, he established links between the distribution of distinctive forms of U. minor and the narrow occurrence of a particular kind of coinage used among ancient tribes in northern France and southern England. He also paid much attention to the period of time when places received a name meaning "elm" (indo-european "*lm-", as in "elm", "ulm", "olm", "alm"…and "*wig-", as in "wice", "wych").

Breeding.

In Europe, elm breeding was mainly achieved in the Netherlands. The Dutch breeding program was continued without interruption over a period of 64 years (1928-1992) and was headed by three persons only (Heybroek, 1993a). In the first phase of the program, elm clones were collected in different European countries, and exotic material was introduced from Asian and American sources. Next phases involved screening for tolerance to DED, crossing selected parents, selecting in the offspring and field testing. Among the hybrid clones obtained by Heybroek, several were released for trade in the Netherlands and others are about to be released in other European countries. Elm breeding has also been undertaken in Italy (Mittempergher et al., 1998) and Spain (Solla, 1999a, b) but these programs are younger than the Dutch one ; one hybrid clone obtained by Mittempergher has been released and others are being field-tested. Besides, screening for tolerance to DED in local material is under way in many European countries ; common protocols for inoculation and symptoms scoring were recently defined in the frame of the EU RESGEN 78 project. A review of the breeding/screening activity carried out in central and eastern European countries would be very helpful when little information is available in the international literature.

The development of biotechnologies and genomics has deeply changed perspectives in breeding work. It is too early to assess the consequences of this change in the case of elms. Several teams have undertaken the genetic transformation of elms, and important progress was made recently (Gartland et al., 1998, 1999).

Gene conservation.

Due to the dramatic consequences of DED pandemics, many European countries have allocated important efforts to the ex situ conservation of their native elm material. This activity is principally built upon vegetative propagation (by grafting and cuttings) and aims at the creation of clonal banks and/or grafted seed-orchards. It is often associated with improvement goals such as selection for tolerance to DED and good phenotype. Conservation of seed lots and ex situ planting of progenies has also been carried out, but to a much smaller extent. In addition, the cryopreservation of elm explants in liquid nitrogen has successfully been experimented in the prospect of long term static conservation purposes (Pâques, 1997). However, the different national programs were defined without any coordination at an international level, and no common protocol was available for the characterization, evaluation, rationalization and long-term management of the collections. Fortunately, the creation of the ‘Noble Hardwoods’ network in 1996 and the launching of the RESGEN 78 EU project in 1997 have reversed this trend (Collin et al., 1999). The 5 year project supported by the European Union is coordinated by Cemagref (France) and involves 17 institutes in 9 west European countries. It is based on the large clone collections existing in several countries and complements them with material originating from partner countries where conservation actions have not yet begun. The clones in these collections will be registered in a common database and characterized with molecular markers applied to nuclear and chloroplast DNA. Their conservation will be optimized in the framework of a European core collection reflecting the geographic partitioning of genetic variation. Conservation methods will associate sustainable field clonal banks (low hedges unattractive for the vector of DED) and cryopreservation of buds in liquid nitrogen. Evaluation will consist of clonal screening for low susceptibility to DED (tolerance to the fungal agent and/or to the insect vector) and good horticultural traits. Despite the fact that this project aims at the rationalization of existing clone collections, and not at the completion of genetic studies based on an independent sampling strategy, it should bring out genetic knowledge of interest for the conservation elm genetic resources throughout Europe.

In the case of U. laevis and some natural stands of U. minor, in situ conservation is often associated with the preservation of riparian ecosystems. In the case of U. glabra, it is often linked with the selection of phenotypically remarkable individual trees by foresters. Such ‘preservation’ approaches are not necessarily sufficient to capture enough genetic variability in the population and thus maintain its evolutionary potential.

More information on elm conservation activities in different European countries can be found in the national presentations published in the meeting reports of the ‘Noble Hardwoods’ network.

The number of parents effectively contributing offsprings is crucial for the maintainance of genetic diversity and additive variance in a population. If this number becomes too low, genetic drift is likely to cause genetic depauperation and loss of ability to adapt to environmental changes. This is why we need to assess the effect of DED on the effective size of elm populations. As U. laevis is generally not severely damaged by the DED pandemics, and as U. minor is not commonly found in large natural populations reproducing sexually, we shall focus on the situation of U. glabra.

1.1 the distribution and dynamics of DED in Europe

First, we need to scrutinize the different factors acting upon the intensity of the pressure exerted by the disease. It is certainly a misleading simplification to speak about DED as if its causal agent was exactly the same everywhere and for ever. Brasier has shown (for a synthetic presentation, see Brasier, 1999a) that the fungus Ophiostoma novo-ulmi responsible for the second epidemics (ca. 1970) in western Europe was not the same and more aggressive as for the first epidemics (ca. 1920). In addition, two different races (the north-American –NAN- and the east-Euroasian -EAN-) exist in Europe and have followed different routes across the continent. Besides, the epidemic process has its own complex dynamics, with phases of low activity when the populations of elms and beetles have decreased in number, and secondary epidemic phases when the surviving young elms have grown to a size allowing a new start in the DED cycle. Moreover, some kind of biological control of the pathogen may be exerted by mycoviruses referred to as ‘d. factors’ (Brasier, 1999b). The role of the insect vectors (different species of bark beetles, with different requirements in the size of the elm trees where they breed, and different efficiencies in inoculating the pathogen to elms) as well as their adaptation to present day climate and adaptability to climatic changes also need to be taken in consideration. For all these reasons, the gravity of the situation is liable to differ among regions and vary considerably over time.

1.2 the effect of DED on population size, age classes, genetic diversity and effective population size

People tend to react emotionally to the visible effects of DED. For instance, in the case of U. minor, the general public in western Europe in the late 70’s had acquired the feeling that elms were bound to disappear ; fifteen years later, the same people were prone to consider that elms had made a spectacular ‘come-back’ in the landscape and that DED was no more a problem. The first feeling and the second consideration are equally wrong and only prove our common failure to estimate properly the excellent regeneration capability of U. minor and the cyclic return of epidemic phases of DED.

Are today’s forest managers in a better position to describe objectively the present situation of U. glabra? When no precise inventories exist, can they totally resist the fallacious impressions they may get from a limited number of observations made by chance (e.g. coming across abundant regeneration of some elm trees in a particular area of the forest)? In the absence of genetic data (e.g. studies with isoenzymes), no one can monitor possible changes in the genetic diversity in the stand after a large proportion of the mature elms are lost to DED? Besides, the desequilibrium in age classes resulting from losses to DED may also considerably decrease the buffer effect that was provided by overlapping generations. Even if an apparently sufficient number of trees are left, is it reasonable to suppose that most of them will survive and equally contribute offspring? Such questions are not trivial when one thinks that the mean effective population size (Ne) to be considered over several generations is not the arithmetic mean of the Ne in each generation, but their harmonic mean "which is always less than the arithmetic mean because it is particularly sensitive to the smallest number of the set" (Lawrence and Marshall, 1997). Thus, special care must be taken to avoid catastrophic reductions in population sizes.

There are very few examples of epidemics causing similar damage and threat to a forest tree species. In the case of sugar pine (Pinus lambertiana), a northern-American species endangered by blister-rust (Cronartium ribicola) introduced from Europe in 1910, Millar (1999) has highlighted the paradox that "a large portion of genetic diversity in sugar pine populations will be lost even though the population recover in size". This paradox is due to a strong selection pressure in favor of a major gene for resistance to the pathogen. To counteract this trend, Millar has recommended to keep the diseased trees (many take years to die) in the stands, such that they "contribute genes through pollen to resistant mother trees" and "also can bear resistant seeds themselves through pollination from resistant father trees". The situation with elms is probably different from the situation with sugar pine because, as observed by Heybroek (1999), full resistance in elms "is apparently determined by many small genes" and "had be gleaned from many sources and accumulated over generations of breeding and selecting in the Dutch breeding program". In the case of the American chestnut (Castanea dentata), a species of the eastern United States devastated by chestnut blight (Endothia parasitica) introduced in 1906, Hamrick and Godt (1996) indicate that no resistance has been found in the species.

In the case of elms and DED, diseased trees generally die in less than 2 or 3 years, so it is unlikely that they contribute seedlings between the moment they show DED symptoms and the moment they die. This situation is particularly bad for U. glabra, which lacks the excellent capability of U. minor for producing root-suckers and re-sprouting from the stump. Moreover, the forest manager cannot really predict when a particular stand will be swept by an active epidemic phase, nor how many trees will remain unaffected, nor why some remain unaffected and others die. Individual elms which are scattered in the forest are more likely to be unnoticed by the insect vectors. It is also possible that others remain uncontaminated just by chance, or because they are unattractive to Scolytus sp. for some unknown reasons (smell ; aspect and taste of bark) or because they have particular characteristics (phenological lag unfavorable to the development of the inoculum in the sap flow ; small diameter of vessels ; induced resistance resulting from another pathologic experience) which enable them to resist a moderate contamination. For all these reasons, it is not possible to predict how many trees will contribute seedlings, nor whether some long lasting individuals will regularly produce hudge quantities of pollen and thus be over-represented in terms of genetic contribution. In order to maintain Ne as high as possible, it would be wise to act preventively, i.e. ensure the natural regeneration of a large number of mature elms before an active epidemic phase occurs or reappears. This precaution is of particular interest in stands that have not yet been severely hit by the epidemics. The difficulty in this task is that it requires a careful silviculture looking after single trees or small patches of trees. There is also a risk that, by giving them more light, the seed trees become more visible to Scolytus sp. ; this risk can be viewed as another reason for undertaking natural regeneration in the stand well before the development of a local epidemic episode.

As the overall state of effective population size in forest elm stands may be considered ‘uncertain’ rather than clearly good or bad (in particular in the case of U. glabra), two courses of action are needed. One is: develop tools to monitor changes, i.e. make periodic inventories in several populations in different regions of Europe, and characterize the genetic diversity in some of them. The other is: take precautionary measures in stand management, i.e. facilitate flowering, seed production and germination, and protect seedlings. These general recommendations do not apply in the case of small marginal populations where urgent and more intensive ex situ and in situ conservation measures may be needed to preserve a rare resource. For budget and continuity reasons, inventories and genetic studies are more likely to be undertaken in forests depending from forestry schools or research institutes. Care for elm regeneration requires a precise silivicultural management and induces additional costs (e.g. soil preparation for seed germination ; protection and possibly fencing of seed patches), but it can certainly be implemented in some state-owned forests or by private owners who have a special interest in elm conservation.

For U. laevis, the major cause of endangerment results probably not from DED but from habitat destruction. As the natural habitat of this species is riparian forests, and as the remaining stands of U. laevis are generally small (e.g. a few ha and no more than 20 mature trees), one can wonder whether in situ conservation is really feasible in such cases due to severe risks of genetic drift. However, such remaining stands are often not very distant (1-10 km) from other populations or even isolated individuals belonging to the same species. What we need to know is whether gene flow occurs between such populations, and whether it is likely to counterbalance the effect of genetic drift. Gene flow may be realized through pollen clouds and/or seeds carried by the stream when the riversides are flooded.

The best way to answer the question is to employ molecular markers and assess the genetic diversity within and between populations along the river. In addition, a detailed analysis of the mating pattern in a population would provide valuable general guidelines for the in situ conservation in similar cases.

A practical manner of dealing with the problem in the absence of genetic knowledge would be to reinforce the population in size and in genetic diversity by planting material (seedlings or clones) obtained from populations in the same fluvial basin. Such reinforcement can be viewed as reintroduction of autochtonous material and reestablishment of continuity between relicts of a formerly unique entity, which was fragmented due to human impact. The preservation and reinforcement of several ‘micro-populations’ along the river may also provide corridors for long term gene flow in a stepping stone effect.

Another and more radical approach would be to consider that, due to genetic drift, the genetic diversity and evolutionary adaptability in the populations have become too low to ensure their long term conservation in situ. The solution proposed would be to sample trees over the whole ecogeographic area, propagate them vegetatively (by grafts or cuttings) and pool their clonal copies in an artificial breeding population. Then, two options are available. One is: create a seed orchard which will ultimately produce the planting material to be reintroduced in the original stands. The other is: create immediately a pseudo-in situ plantation with plantlets obtained by cuttings (i.e. grown on their own roots, as opposed to grafted plantlets) and planted in the same area and kind of environment as the original stands. In the first option, management is carried out ex situ and artificially, and no output is immediately available in situ. In the second option, pseudo-in situ plantation is undertaken two years only after the collection of cuttings, and the stand can be further managed just like a natural stand (i.e. allowing trees to grow to their full size and be submitted to natural selection and silvicultural thinnings). The goals and by-products of these two conservation schemes are not identical: in the first case, seed production is clearly targeted and may provide a commercial output ; in the second case, immediate conservation in the original habitat is emphasized, while seed production becomes optional. Conversely, conservation of associated bio-diversity and landscape value may be not negligible.

The strongest consequence of human impact on the resource of U. minor is undoubtedly the catastrophic damage due to the introduction of a new pathogen in Europe! Paradoxically, the former effects of human activities were extremely beneficial to U. minor, which had been extensively propagated and planted (some may say ‘overplanted’) in many countries and many kinds of environments (including urban).

What we need to assess now is: i) what are the present consequences of ancient plantings on the geographic partitioning of genetic variation in U. minor ii) what are the possible effects of genetic improvement programs on the gene pool of U. minor.

3.1 is there an architecture behind chaos?

As discussed above, the notion of population is not always easy to delineate practically in the case of sexually reproducing ‘wild’ species of elms such as U. laevis and U. glabra. In the case of the suckering ‘semi-cultivated’ and taxonomically complex Field elms, the definition of populations may become extremely awkward, if not hopeless. Long distance transfers of planting material by humans, combined with natural vegetative propagation in hedgerows have probably produced a complex pattern of variation, eventually complicated by subsequent hybridization between varieties of U. minor and/or introgression by U. glabra. The occurrence of sexual reproduction in U. minor varies considerably according to regional climate. In Great-Britain, it is so rare that populations of Field elms can be regarded as principally composed of a diverse array of clonal material. In southern Europe, U. minor sets seed abundantly and frequently, but the notion of population is still complicated by the effect of natural suckering and human management.

This situation may lead to broaden the notion of ‘population’ when defining geographical units for conservation activities or the computation of among-units differentiation. In France, the unit scale pragmatically chosen for such activities and computations was the size of a large geographical region (Machon, 1997). This choice may be rejected from a biological point of view ; it can also be regarded as an acceptable form of stratified sampling based on historic and ecogeographic entities. Many other alternatives are available, and different situations in different countries may lead to completely different approaches. In particular, if some riparian or island populations have not been subject to extensive anthropogenic management, their conservation could be treated in a manner very similar to U. laevis and U. glabra, provided that special attention is paid to the presence of clonal patches in the stand.

3.2 use of exotic species, hybrid cultivars and transgenic material

The debate about the introduction of ‘alien’ gene resources in agriculture is both technical and ethical. In the case of elms, it would be relevant to collect all kinds of data of interest in the technical part of the debate. For instance, isozyme studies are carried out in Spain to recognize and assess the introgression of local U. minor by the introduced Asian species U. pumila. Ethical debates about ‘genetic pollution’ and discussions about the benefits (improved tolerance to DED) and dangers (deterioration of wood quality) of such introgression may be viewed as mere ‘exercises of style’ as long as no quantitative data are available to describe the reality and possible evolution of this phenomenon. The same kind of concern exists in the case of hybrid material released for trade in several European countries. Cultivars obtained from crosses of Asian species (e.g. ‘Sapporo Autumn Gold’) are widely planted, not only in cities but also for the reconstruction of countryside hedges, and are likely to shed large quantities of pollen. Before debating about the consequences of this situation, it would be necessary to gather precise data on such plantations and the possible lag between the flowering periods of those cultivars and the local elms.

Genetic engineering and risks of ‘gene escapes’ are issues that may arise in a near future in the case of elms. A clone of English elm (U. minor var. vulgaris, syn. U. procera) is being genetically modified for resistance to DED (Gartland et al., 1999). One argument that may be used in favor of its possible release is that, at least in Great Britain, U. procera does not flower, or very seldom, which provides a certain guarantee against ‘gene escapes’.

In the case of elms, regulations about the use of cultivars are mainly taken in the scope of urban plantings. They can be regarded as maladapted to plantings in other situations (forestry, reconstitution of countryside hedgerows,…).

As indicated by Eriksson 1996 and other authors (Brown and Briggs, 1991), alleles that occur in intermediate frequencies (say 10-90 %) are the most interesting in an evolutionary perspective. According to the logarithmic relationship expressed by Brown, there is "a strong law of diminishing (genetic) returns on sample size in terms of the allele richness of the sample. The first ten organisms randomly sampled from a population are as important as, if not more important than, the additional 90" (Brown and Briggs, 1991). This law of ‘diminishing returns’ suggests that we devote ex situ collection efforts to taking few samples in many populations in different ecogeographic zones rather many samples in few populations.

In addition to alleles randomly sampled, should we also care about genotypes, or ‘genets’, deliberately chosen ? This is not justified if conserving additive variance is our sole goal but it might be of interest to capture genets when some kind of tolerance to DED or other valuable traits are likely to be gleaned. In the case of mature elms that have survived the epidemics, such captures will be vain if the tree survived just by chance, but they may also bring out a character of interest in selection and breeding work.

Ex situ conservation must be regarded as complementary to in situ conservation. In the case of elms, it will mainly consist in the preservation of clonal copies (obtained using cuttings or grafts) of genotypes, and be employed when emergency measures are needed for rare endangered populations (e.g. U. glabra in Sicily) and when populations are too small to be managed in situ (e.g. risks of genetic drift). It will also be convenient for U. minor, which does generally not form real ‘populations’ but complex mixtures of clonal populations in which taxonomic diversity and tolerance to DED can be targeted using stratified sampling methods. In addition, ex situ conservation is an excellent means for building collections easily available for research work.

What needs to be defined is whether ex situ measures should also be applied for populations in which in situ conservation is carried out, and whether there should be many more ex situ than in situ sites. Millar and Libby (1991) give positive answers to both questions, such that clinal and ecotypic diversity can be properly sampled. As one may doubt that enough money and manpower will be available for such an extensive sampling, an alternative solution would be to schedule the ‘two-stage sampling’ scheme proposed by Jain (cited in Brown, 1992), "where diversity studies of the first round of samples are used to improve the second round".

A large amount of clonal material and genetic knowledge has already been accumulated in the frame of past conservation and/or breeding programs, and it is important not to loose it when people retire and/or institute policies change. Heybroek (1999) has witnessed such difficulties, and other experiences show that ‘orphan’ clone collections become very fragile. International projects and networks provide good opportunities for reviving old collections and facilitating their characterization and use. For instance, the RESGEN78 EU project has enabled the duplication of the material collected by Melville in the 70’s and stimulated its use for research on U. plotii. Several ex situ collections of elm clones in Europe are probably in great need of inventory and rejuvenation.

Foresters and forest geneticists must develop synergies with other specialists of elms (botanists, breeders, pathologists and epidemiologists,…) as well with conservationists and elm enthusiasts. For instance, conservationists who take care of natural reserves along large rivers can play a major role in the discovery, inventory, preservation and regeneration of U. laevis. Raising awareness is a major issue, especially among conservationists who may be more interested in habitat preservation than in the dynamic management of infra-specific gene resources. Exchanges of information among managers of gene conservation stands is crucial for the implementation of efficient regeneration methods ; dissemination of information can be facilitated by the Noble Hardwoods network, but direct exchanges among managers through the Internet needs to be encouraged. A Website should be made available such that managers of sites not officially selected by the network can be associated in the conservation effort and informed of progress and experience in other places.

The different issues discussed above show that research, breeding, conservation and forest management should not be considered independently but that progress in one of these activities enables some progress in the other fields as well. The link between research and conservation activities should particularly be emphasized, such that sampling for conservation and genetic studies are coordinated and reinforce each other. The same situation applies in terms of geographic scale. Gene conservation programs can be undertaken separately in different countries, but they are certainly more efficient when coordinated. For instance, if research and conservation work is to be undertaken on the gene resources of U. laevis in the Rhine valley or the Danube basin, it is needed that sampling be organized in the different countries along the rivers and that genetic analyses are carried out with the same protocol. Defining comprehensive research and conservation strategies, and providing help for coordination of national projects are two major tasks of the EUFORGEN Noble Hardwork network.

Suggested activities to be carried out in the network are:

- contribution to the rationalization of sampling throughout the distribution range

- holding record of all relevant data on conservation units and persons involved

- identification of research fields of major importance

- contribution to the preparation of proposals for research and conservation programs

- making recommendations as regards the regulations on the transfer of planting material (provenance of forest reproductive material ; use of ornamental cultivars in natural areas)

 

General indications for sampling for in situ conservation of common Noble Hardwoods have been discussed by Jensen et al. (1999): select at least 30 populations/species over the distribution area ; effective size of each population (Ne) must be greater than 50; in the center of the distribution area, give preference to large gene reserves(Ne>150) ; in marginal areas, select a larger number of conservation units of smaller size.

Additional remarks for elms are as follows:

· sampling for in situ conservation should be stratified and based upon environmental factors (ecogeographic zones, habitats) and genetic factors (subspecies, varieties, ecotypes, major genetic differenciation revealed with markers) ; special attention should be given to populations at the southern margin of the distribution area

· requirements as regards the number of populations/species should be put up for U. minor and U. glabra because of the occurrence of regional forms of high taxonomic rank (variety, sub-species or even ‘species’)

· requirements as regards the effective size of population (Ne = 50) should be interpreted according to risks of losses (due to DED) and possibilities for gene flow

· emergency measures should be rapidly taken for the preservation of endangered rare populations (e.g. risks of habitat destruction or maladapted management)

· in all other cases (i.e. larger and less threatened populations), conservation units should be selected only when all relevant documents and inventories are available ; in addition to environmental factors and genetic factors, selection criteria should include guaranties for a sustainable and precise management ; preference should be given to stands where joint conservation with another species in the EUFORGEN programme is possible (e.g. U. laevis with Populus nigra or Alnus sp. in riparian forests ; U. glabra with Acer sp. or Tilia sp. in montane stands) ; attention should also be paid to associated biodiversity (e.g. rare species of lichens in the case of U. glabra)

· the notion of population might be maladapted to U. minor due to anthropic management and/or extensive vegetative propagation ; in such cases, ex situ conservation may be preferred to in situ

General indications for in situ management have also been given by Jensen et al. (1999): stimulate natural regeneration, plant local material when natural regeneration is not sufficient, avoid contamination with pollen from external sources, practice weeding, clearing and thinning. Additional remarks for elms are as follows:

· in the case of small fragmented populations of U. laevis, reinforce the size and diversity of the population by planting material obtained from the same fluvial basin and environmental conditions

· in the case of U. glabra, stimulate natural regeneration as a precautionary measure, such that seedlings are in place before the mother trees become diseased

· in the case of U. minor, cut the diseased trees and protect their re-sprouts, stimulate root-suckering if necessary

· when an active epidemic phase arises, practice sanitation (prophylactic cuttings and pruning, trap and destroy bark beetles,…) in the conservation unit and in its vicinity.

 

General indications for sampling for ex situ conservation of common Noble Hardwoods have been discussed by Jensen et al. (1999) and above (under 4).

Additional remarks for elms are as follows:

· ex situ conservation will mainly - but not only - be employed when population size is too small to allow in situ conservation ; it is an excellent tool for the conservation of U. minor, which is highly variable and shows very complex patterns of geographic variation ; ex situ conservation will also be convenient to capture genotypes of interest for research and breeding work

· sampling for ex situ conservation should be stratified and based upon environmental factors (ecogeographic zones, habitats) and genetic factors (subspecies, varieties, ecotypes, major genetic differenciation revealed with markers) ; special attention should be given to populations at the margins (in particular the southern one) of the distribution area as well as to individual trees that may own some tolerance to DED

· the number of populations per species where ex situ conservation should be undertaken depends on whether ex situ measures can take place in units where in situ conservation is already being carried out ; if funds allocated enable such a complementary approach, ex situ conservation will be used for building collections easily available for research and breeding work

· preference will be given to the collection of cuttings or grafts, but seed and seedlings can be used as an alternative solution

· the number of individuals to sample per population should be about 10 ; a higher number can be sampled if this is not detrimental (in terms of funds and manpower available) to sampling in another population in a different ecogeographic zone

· emergency measures should be rapidly taken for the preservation of endangered rare populations (e.g. mortality due to DED, risks of habitat destruction)

· sampling for allelic diversity within a stand should be carried out randomly ; in addition, sampling for valuable genets can be specifically targeted in the case of trees with a remarkable phenotype or which have survived a very severe epidemic phase of DED ; trees with particular botanic or phenologic traits should be sampled too

· a minimum distance of 50 m between trees to sample is needed in the case of U. minor, which often forms clonal patches (this precaution is not needed when phenotypic or phenologic differences are obvious)

· material to be made available for research and breeding work should be preserved in collections built and managed as low hedges (less attractive to insect vectors of DED) ; a small subset of the collection can be cryopreserved (buds) if such facility is available at a reasonable cost and with good chances of subsequent regeneration ; existing collections should be described (and characterized if possible) and rationalized to ensure their conservation in the long term (duplication, constitution of core collections)

· as suggested by J. Kleinschmit (pers. com.), material for seed production should be assembled in grafted seed-orchards comprising at least 30 genotypes from the same ecogeographic zone ; size of orchard should be at least 1 ha ; spacing 5 x 5 m with a possibility to enlarge it later to 5 x 10 m)

· material for pseudo-in situ conservation should be obtained from seed and/or cuttings and be composed of at least 30 genotypes or 20 progenies from the same small ecogeographic zone (e.g. fragmented populations belonging to the same riparian ecosystem) ; after plantation in the original habitat, management is as indicated for in situ conservation

 

Two different and complementary courses of actions are urgently needed: i) collect and improve genetic knowledge on elms ; ii) implement and rationalize conservation activities in a pan-European perspective. Such actions need to be decided and carried out at country level, but it is the task of the EUFORGEN Noble Hardwoods network to identify need and indicate priorities.

In spite of today’s facilities for exchanges of information, existing knowledge is not always fully valorized. Translations and literature reviews are necessary, and can be encouraged by the network. Research fields of major importance must be highlighted, and submission of projects in international calls for proposals facilitated. Applied research, such as the development of resource inventories and monitoring of changes in genetic diversity, should be promoted too. Independently from the network, meetings such as ‘The International Elm Conference’ are likely to provide excellent opportunities for collecting and exchanging information. In addition, results from diverse ongoing research projects should also become available in the next few years

It is the role of the network to contribute to the rationalization of sampling throughout the distribution range and to hold record of all relevant general data on conservation units and contact persons or services. Besides, links with other EUFORGEN networks (in particular ‘Populus nigra’) should be encouraged for the development of a common ‘habitat’ approach. Awareness must also be risen through publications and the Internet.

At country level, research and conservation activities must necessarily be a compromise between the need already identified and the funds allocated to the projects. Priorities should be given to preserving endangered natural populations and existing ex situ collections (if any). Inventories and pre-selection of conservation units should come next, together with raising awareness about genetic conservation methods among forest managers and environmentalist organizations. If possible, this preliminary phase should include studies of genetic variation. Finally, long term conservation management should be ensured in carefully selected populations, representative of a significant portion of elm diversity in the country and from a European perspective.

Acknowledgments: appreciation is greatly expressed to the many people, and particularly Drs. S. Wiegrefe and N. Machon, who have facilitated my access to literature resources.

Armstrong, J. V. 1992. Taxonomy of British elms. PhD. University of Cambridge.
Armstrong, J. V., and P. D. Sell. 1996. A revision of the British elms (Ulmus L., Ulmaceae): the historical background. Botanical Journal of the Linnean Society 120: 39-50.
Bate Smith, E. C., and R. H. Richens. 1973. Flavonoid chemistry and taxonomy in Ulmus. Biochemical Systematics 1: 141-146.
Borlea, G. F. 1995. The elm species in Rumania. Pages 407-411 in P. Baradat, W. T. Adams, and G. Muller-Starck, eds. Population genetics and genetic conservation of forest trees. SPB Academic Publishing, Amsterdam, The Netherlands.
Brasier, C. M. 2000a. Intercontinental spread and continuing evolution of the Dutch elm disease pathogenes. Pages 61-72 in C. P. Dunn, ed. The Elms: breeding, conservation and disease management. Kluwer Academic Publishers, Boston, Mass., USA.
Brasier, C. M. 2000b. Viruses as biological control agents of the Dutch elm disease fungus Ophiostoma novo-ulmi. Pages 201-212 in C. P. Dunn, ed. The Elms: breeding, conservation and disease management. Kluwer Academic Publishers, Boston, Mass., USA.
Brown, A. H. D. 1992. Human impact on plant gene pools and sampling for their conservation. Oikos 63: 109-118.
Brown, A. H. D., and J. D. Briggs. 1991. Sampling strategies for genetic variation in ex situ collections of endangered plant species. Pages 99-119 in D. A. Falk and K. E. Holsinger, eds. Genetics and conservation of rare plants. Oxford University Press, New York.
Coleman, M. 1998. The elm problem: a molecular and morphological case study. M.Sc. degree in Plant Taxonomy and Biodiversity. Edinburgh University, and the Royal Botanic Garden Edinburgh.
Coleman, M., M. L. Hollingsworth, and P. M. Hollingsworth. in press. Application of RAPDs to the critical taxonomy of the English endemic Ulmus plotii Druce. Botanical Journal of the Linnean Society .
Collin, E., I. Bilger, G. Eriksson, and J. Turok. 2000. The conservation of elm genetic resources in Europe. Pages 281-293 in C. P. Dunn, ed. The Elms: breeding, conservation and disease management. Kluwer Academic Publishers, Boston, Mass., USA.
Daumann, E. 1975. Ein rudimentares Blutennektarium und unterschiedliche Pollenkittreste bei der Ruster (Ulmus). [A rudimentary floral nectary and various pollen-adhesive residues in the elm (Ulmus)]. Preslia (Prague) 47: 14-21.
Eriksson, G. 1996. Evolutionary genetics and conservation of forest tree genetic resources. Pages 159-167 in J. Turok, G. Eriksson, J. Kleinschmit, and S. Canger, eds. EUFORGEN Noble Hardwoods Network: report of first meeting, Escherode, Germany, 24-27 March 1996. International Plant Genetic Resources Institute, Rome, Italy.
Gartland, J. S., T. M. Fenning, C. M. Brasier, A. T. McHugh, K. M. A. Gartland, M. R. Davey, P. G. Alderson, K. C. Lowe, and J. B. Power. 1998. Transformation studies with Ulmus species. Pages 285-290. Tree biotechnology: towards the millennium. Nottingham University Press, Nottingham.
Gartland, K. M. A., J. S. Gartland, T. M. Fenning, A. T. McHugh, R. J. Irvine, G. D. Main, and C. M. Brasier. 2000. Genetic manipulations with elms. Pages 259-270 in C. P. Dunn, ed. The Elms: breeding, conservation and disease management. Kluwer Academic Publishers, Boston, Mass., USA.
Hamrick, J. L., and M. J. Godt. 1996. Conservation genetics of endemic plant species. Pages 281-304 in J. C. Avise and J. L. Hamrick, eds. Conservation genetics: case histories from nature. Chapman and Hall, New York (USA).
Heimler, D., L. Mittempergher, P. Buzzini, and V. Boddi. 1990. Quantitative HPTLC separation of flavonoid glycosides in the taxonomy of elm (Ulmus spp.). Chromatographia 29: 16-20.
Heimler, D., A. Pieroni, L. Mittempergher, and P. Buzzini. 1993. The use of flavonoid glycosides for the identification of elm hybrids. Canadian Journal of Forest Research 23: 611-616.
Heybroek, H. M. 1990. Los olmos, compañeros de la humanidad. [The elms, companions of Man]. Pages 17-25 in L. Gil, ed. Los olmos y la grafiosis en España. [The elms and Dutch elm disease in Spain]. ICONA, Madrid.
Heybroek, H. M. 1993a. The Dutch elm breeding program. Pages 16-25 in M. B. Sticklen and J. L. Sherald, eds. Dutch elm disease research: cellular and molecular approaches. Springer-Verlag, New York (USA).
Heybroek, H. M. 1993b. Why bother about the elm? Pages 1-8 in M. B. Sticklen and J. L. Sherald, eds. Dutch elm disease research: cellular and molecular approaches. Springer-Verlag, New York (USA).
Heybroek, H. M. 2000. Notes on elm breeding and genetics. Pages 249-258 in C. P. Dunn, ed. The Elms: breeding, conservation and disease management. Kluwer Academic Publishers, Boston, Mass., USA.
Hollingsworth, P. M., M. L. Hollingsworth, and M. Coleman. 2000. The European elms: molecular markers, population genetics, and biosystematics. Pages 3-20 in C. P. Dunn, ed. The Elms: breeding, conservation and disease management. Kluwer Academic Publishers, Boston, Mass., USA.
Ipinza, R. 1989. Contribucion a la taxonomia numerica de los olmos ibericos a través de algoritmos de classificacion jerarquica. [Contribution to the numeric taxonomy of the Spanish elms using hierarchical classification algorithms]. PhD thesis. University of Madrid.
Jeffers, J. N. R. 1999. Leaf variation in the genus Ulmus. Forestry 72: 183-190.
Jeffers, J. N. R., and R. H. Richens. 1970. Multivariate analysis of the English Elm population. Silvae Genet. 19: 31-38.
Jensen, J. S., M. Rusanen, and A. Pliura. 1999. Gene conservation of Noble Hardwoods in situations where they occur widespread. Background document for the Fourth meeting of the EUFORGEN Noble Hardwoods Network, Gmunden, Austria, 3-7 Sept. 1999. 10 p.
Jobling, J., and A. F. Mitchell. 1974. Field recognition of British Elms. HMSO. Forestry Commission Booklet n° 42, London.
Lawrence, M. J., and D. F. Marshall. 1997. Plant population genetics. Pages 99-113 in N. Maxted, B. V. Ford-Llyod, and J. G. Hawkes, eds. Plant Genetic Conservation. Chapman & Hall.
Machon, N., M. Lefranc, I. Bilger, and J. P. Henry. 1995. Isoenzymes as an aid to clarify the taxonomy of French elms. Heredity 74: 39-47.
Machon, N., M. Lefranc, I. Bilger, S. J. Mazer, and A. Sarr. 1997. Allozyme variation in Ulmus species from France: analysis of differentiation. Heredity 78: 12-20.
Mattila, A., and P. Vakkari. 1997. Genetic variation of Quercus robur and Ulmus laevis in Finland. Pages 63-68 in M. Kurm and U. Tamm, eds. Metsa geneetiliste ressursside sailitamine. Conservation of forest genetic resources. Eesti Pollumajandusulikool. Estonian Agricultural University, Tartu.
Melville, R. 1938. Is Ulmus campestris L. a nomen ambiguum? J. Bot. : 261-265.
Melville, R. 1940. Contributions to the study of British Elms. III. The Plot Elm, Ulmus plotii Druce. J. Bot. 78: 181-192.
Melville, R. 1957. Ulmus canescens: an eastern Mediterranean Elm. Kew Bull. : 499-502.
Melville, R. 1975. Ulmus. Pages 292-299 in C. E. Stace, ed. Hybridization and the flora of the British Isles. Academic Press, London.
Melville, R., and R. H. Richens. 1978. On the discrimination of species in hybrid swarms with special reference to Ulmus and the nomenclature of U. minor Mill. and U. carpinifolia Gled. Taxon 27: 345-351.
Millar, C. I. 1999. Genetic diversity. Pages 460-494 in M. L. J. Hunter, ed. Maintaining Biodiversity in Forest Ecosystems. Cambridge University Press, Cambridge, UK.
Millar, C. I., and W. J. Libby. 1991. Strategies for conserving clinal, ecotypic, and disjunct population diversity in widespread species. Pages 149-170 in D. A. Falk and K. E. Holsinger, eds. Genetics and conservation of rare plants. Oxford University Press, New York.
Mittempergher, L., A. Fagnani, and F. Ferrini. 1998. Gli olmi e la grafiosi in città oggi. [Elms and Dutch elm disease in town today]. Informatore fitopatologico 6: 34-39.
Mittempergher, L., and N. La Porta. 1991. Hybridization studies in the Eurasian species of elm (Ulmus spp.). Silvae Genetica 40: 237-243.
Paques, M., M. Poissonnier, E. Dumas, V. Monod, A. Altman, and M. Ziv. 1997. Cryopreservation of dormant and non-dormant broad-leaved trees. Proceedings of the third international ISHS symposium on in vitro culture and horticultural breeding, Jerusalem, Israel, 16 21 June, 1996. Acta Horticulturae. 1997, No. 447, 491 497.
Richens, R. H. 1963. Monophage analysis of Elm populations. FAO World Consult. For. Genet., Stockholm 1963 No. FAO/FORGEN 63/-6b/4.
Richens, R. H. 1968. The correct designation of the European field elm. Feddes Repertorium : 1-2.
Richens, R. H. 1976. Variation, cytogenetics and breeding of the European field elm (Ulmus minor Miller sensu latissimo = U. carpinifolia Suckow). Anali za Sumarstvo. Annales Forestales (Zagreb) 7: 107-145.
Richens, R. H. 1980. On fine distinctions in Ulmus L. Taxon 29: 305-312.
Richens, R. H. 1983. Elm. Cambridge University Press, Cambridge.
Richens, R. H., and J. N. R. Jeffers. 1975. Multivariate analysis of the elms of northern France. I. Variation within France. Silvae Genetica 25: 141-150.
Richens, R. H., and J. N. R. Jeffers. 1978. Multivariate analysis of the elms of northern France. II. Pooled analysis of the elm populations of northern France and England. Silvae Genetica 27: 85-95.
Richens, R. H., and J. N. R. Jeffers. 1985. The elms of Wales. Forestry 58: 8-25.
Richens, R. H., and J. N. R. Jeffers. 1986. Numerical taxonomy and ethnobotany of the elms of northern Spain. An. Jard. Bot. Madr. 42: 325-341.
Sacchetti, P., R. Tiberi, and L. Mittempergher. 1990. Preference of Scolytus multistriatus (Marsham) during the gonad maturation phase between two species of elm. Redia 73: 347-354.
Solla, A., Y. Menendez, M. Buron, S. Iglesias, and L. Gil. 2000a. Programa español para la conservacion y mejora de Ulmus minor Miller. [Spanish program for the conservation and improvement of Ulmus minor Miller]. Montes : 5-12.
Solla, A., Y. Menendez, M. Buron, S. Iglesias, and L. Gil. 2000b. Spanish program for the conservation and breeding of elms against DED. Pages 295-304 in C. P. Dunn, ed. The Elms: breeding, conservation and disease management. Kluwer Academic Publishers, Boston, Mass., USA.
Stockmarr, J. 1970. Species identification of Ulmus pollen. Danmarks Geologiske Undersoegelse, IV 4: 1-19.
Stockmarr, J. 1974. Sem studies on pollen grains of North European Ulmus species. Grana 14: 103-107.
Timbal, J., and E. Collin. in press. L'orme lisse (Ulmus laevis Pall.) dans le sud de la France: répartition et stratégie de conservation des ressources génétiques. [The European White Elm (Ulmus laevis Pall.) in southern France: distribution and genetic resources conservation strategy]. Revue Forestière Française .
Uotila, P. 1997. Jalavan suku sekä lehto- ja vuorijalavan taksonomia. [The genus Ulmus, and the taxonomy of U. minor and U. glabra]. Sorbifolia 28: 5-16.
Walter, H. 1931. Ulmaceen. Pages 639-719 in O. V. Kirchner, E. Loew, and C. Schröter, eds. Lebensgeschichte der Blütenpflanzen Deutschlands, Leipzig (Germany).
Webber, J. 2000. Insect vector behavior and the evolution of Dutch elm disease. Pages 47-60 in C. P. Dunn, ed. The Elms: breeding, conservation and disease management. Kluwer Academic Publishers, Boston, Mass., USA.
Wiegrefe, S. J. 1992. Molecular genetic variation in the Ulmaceae: phylogenetic implications. PhD. thesis. University of Wisconsin-Madison.
Wiegrefe, S. J., R. P. Guries, E. B. Smalley, and K. J. Sytsma. 1993. Genetic diversity in elms: what molecular data tell us. Pages 227-238 in M. B. Sticklen and J. L. Sherald, eds. Dutch elm disease research: cellular and molecular approaches. Springer-Verlag, New York (USA).
Yarmolenko, A. V. 1936. Urticales Lindl. Pages 283-296 in V. L. Komarov, ed. Flora of the USSR, Moskva-Leningrad.