1. Introduction

Despite the efforts of conservationists, biological diversity faces severe threats and continues to be lost at increasing rates around the world (e.g. Pimm et al., 1995; Butchart et al., 2010; Ceballos et al., 2015). To date, plant conservation has focused on passive protection of fragmented natural habitats. This approach, however, appears inadequate to the task of reducing the accelerating losses of both species and natural habitats globally (e.g. Heywood, 2016, 2017; Fenu et al., 2017a). In situ conservation, which is widely accepted as the primary and most appropriate strategy for conserving biodiversity, focuses on populations in their natural habitats (e.g. Godefroid et al., 2011; Volis, 2017; Heywood, 2019). One relatively recently developed in situ activity called plant translocation is a potentially important tool for reducing extinction risks for threatened species and improving the conservation status of these threatened species (e.g. Maschinski and Duquesnel, 2006; Godefroid et al., 2011; Orsenigo, 2018).

Conservation translocation (translocation, hereafter) is the controlled placement of plant material into a (semi-) natural area - managed or not - and includes population reinforcement, reintroduction, and introduction. The aim of translocation is to increase survival of a given species and has thus been encouraged as an approach to prevent the extinction of plant species (e.g. Maschinski and Duquesnel, 2006; Godefroid et al., 2011; IUCN/SSC, 2013; Volis, 2016a; Laguna et al., 2016; Commander et al., 2018). The potential of translocations to contribute to the recovery of threatened species is significant, and is enhanced when done as a part of an integrated conservation plan (Maschinski and Duquesnel, 2006; Albrecht et al., 2011; Cogoni et al., 2013; Volis, 2016b, 2017). Nevertheless, despite the strategic importance of in situ measures as highlighted by the most important international conventions and programmes (e.g. Article 9C of the CBD, Target 7 of the GSPC for 2020), their implementation lags far behind their full potential. The reasons for this are numerous (e.g. Godefroid et al., 2011; IUCN/SSC, 2013; Commander et al., 2018). For example, many translocation studies are not well-known by other scientists and conservation practitioners, making the lessons of translocation successes or failures difficult to learn (Meek et al., 2015; Volis, 2019). Therefore, to improve the knowledge and awareness of translocation studies within the scientific community, it is important to share all experiences from translocations, even when they are not innovative.

The Mediterranean Basin is an important centre of plant diversity. Although it only consists of 1.6% of the Earth's surface, the region hosts approx. 7% of the world's plants, and accordingly has been identified as one of the global biodiversity hotspots (Médail and Quézel, 1997; Cañadas et al., 2014). However, this plant diversity is unevenly distributed (Thompson, 2005; Cañadas et al., 2014); specifically, most plant diversity is located on the big Mediterranean islands (i.e. Sicily, Sardinia, Cyprus, Corsica and Crete) and Balearic Archipelago, which have an endemism rate of more than 40% (Fenu et al., 2017b). In fact, the peculiar Mediterranean insular features determine specific plant diversities and assemblages and, as a consequence, within the Mediterranean Basin, islands and islets constitute the main centers of plant diversity chiefly due to the narrow range of most of their flora (Médail and Quézel, 1997; Vogiatzakis et al., 2016; Fenu et al., 2017b). At the same time, it is also well known that such plant richness is severely threatened by several factors (physical and biological) and, consequently, many plants of these islands require urgent measures, including but not limited to protection.

Plant diversity of the Mediterranean insular territories is correlated with several human activities that have recently had relevant, negative consequences on plant distributions and dynamics. Currently, insular Mediterranean plant diversity is severely threatened by both natural and anthropogenic factors and its conservation deserves particular attention. Over the last four millennia, the Mediterranean Basin has been the cradle of some of the world's greatest civilizations, a situation that has resulted in soil over-exploitation and the conversion of much of the pristine vegetation into agricultural landscapes (Vogiatzakis et al., 2016). Moreover, the Mediterranean Basin is one of the most vulnerable regions to climate changes, exposing the Mediterranean islands to unique challenges (IPCC, 2013; Vogiatzakis et al., 2016; Cramer et al., 2018).

Because of their limited area, discrete nature, and simplified food webs, islands have often been considered "natural laboratories" for ecological studies, including translocation experiments. Moreover, areas like the Mediterranean insular territories are ideal test sites because of both high rates of endemism and a large number of environmental/anthropogenic threats to plant biodiversity. These shared factors provide an opportunity to test the effectiveness of different combinations of methods and methodologies at conserving endangered plants on Mediterranean islands/ islets. Despite the urgent need for translocation actions for a great number of local narrow endemics, few documented translocations have been carried out in the Mediterranean territories (e.g. Piazza et al., 2011; Cogoni et al., 2013; Rita and Cursach, 2013; Heywood, 2014; Laguna et al., 2016). Furthermore, to our knowledge, no translocation projects in this region have focused on methodological aspects.

2. Care-Mediflora approach

The Care-Mediflora project (http://www.care-mediflora.eu/), supported by the MAVA Foundation, is an initiative led by institutions of six Mediterranean islands (mostly botanical gardens and/or seedbanks) and the IUCN/SSC Mediterranean Plant Specialist Group. Participating institutions have a deep knowledge of the local flora and their conservation status, and also have extensive experience with ex situ conservation. All institutions jointly worked to address both short-term and long-term needs for the insular endangered plants, including: 1) in situ conservation through active management actions, in particular translocations, and 2) ex situ conservation through the seed banking of accessions representative of the overall diversity as a tool for plant multiplication for future in situ conservation actions (Fenu et al., 2017b).

The approach agreed upon by all partners is mainly based on the use of ex situ activities and knowledge as a tool to improve in situ conservation of threatened plant species. For example, genetic material (seeds) and know-how from previous ex situ activities (e.g., seed collections, germination experiments, living plant collections, etc.) was used in field work to conserve threatened plants in situ, particularly through translocation programs. This approach could be defined "from ex situ to in situ conservation" and it is consistent with the need to integrate ex situ and in situ approaches, which was recognized more than four decades ago (e.g. Falk, 1987; Heywood, 1993; Maunder et al., 2001). The Care-Mediflora project applies the experience gained from a number of conservation programs that have utilized material maintained and/or propagated in botanic gardens for in situ translocations (e.g. Maschinski and Duquesnel, 2006; Noël et al., 2011; Menges et al., 2016; Fenu et al., 2016). The Care-Mediflora approach can be framed in the conservation-oriented restoration concept (Volis, 2016a, b). Both are grounded on the same two major principles: (1) there are no alternative ways to actively manage populations of threatened species to prevent their extinction; (2) wide-scale plant introductions of threatened species - both within and outside of known historical ranges for the species - is essential.

3. Translocations of the care-Mediflora project

Any translocation requires an in-depth understanding of the biology and ecology of the target species (e.g. life form, reproductive biology, demographic parameters, etc.), as well as the main threats affecting the population (e.g. Maunder et al., 2004; Maschinski and Duquesnel, 2006; Noël et al., 2011; Cogoni et al., 2013; IUCN/SSC, 2013; Commander et al., 2018). The translocation process often requires relevant expert-based decisions about whatand how-to-do site selection, origin and type of the outplants (seeds, seedlings, or cuttings), planting methods, and the appropriate monitoring plan (Maunder et al., 2004; Godefroid et al., 2011; IUCN/SSC, 2013; Commander et al., 2018). Also, the expertise in ex situ multiplication and cultivation procedures is a crucial prerequisite (Cogoni et al., 2013; Brancaleoni et al., 2018).

In the preliminary stage, all the partners jointly established a set of common criteria for selecting target species and prioritizing their conservation actions. In particular, four main criteria were considered crucial: 1) how threatened a species is according to the global/ regional IUCN Red List, 2) the regional responsibility criterion that represents the first order of priority at the local level; 3) the socalled policy plant species criterion, giving priority to the species listed in the annexes of Habitat Directive and/or in other national or regional regulations, and 4) affinity to the wetland habitat, the latter considered most sensitive to the on-going climate change/ climatic instability (Fenu et al., 2017b and references therein).

Based on these common criteria, six regional lists (one per partner) and a global list of target plants in a need of conservation were compiled; then starting from the local list, each partner selected a further group of plants in an urgent need of in situ conservation actions. Following this procedure, a global list of 168 threatened taxa in the six Mediterranean islands was compiled, which includes mainly plants selected by the Regional Responsibility Criterion (79.07%) and plants assessed as threatened in the global and/or regional IUCN Red List (67.99%). Plants listed in the Habitat Directive, national or regional regulations and plants related to wetland areas accounted for 23.26% and 9.30%, respectively. Based on these conservation priorities, each partner worked on a set of target plants according to the available knowledge of a candidate species range, the main threats and biotic/abiotic requirements, as well as the feasibility of carrying out a translocation program. For each candidate species, all the existing data on extant and historical distribution, biology and ecology, and the main threats were examined to verify whether a translocation was the best solution to reduce the species extinction risk.

To date, 50 translocations (involving 45 taxa) have been implemented at six Mediterranean islands using a variety of protocols, which differed in type of translocation, site selection method, origin of the genetic material, type of propagated material (seeds or cuttings), and/or planting method (Tables 1 and 2).

The target species represented different life histories, mostly hemicryptophytes and chamaephytes, and a variety of habitats (ranging from sandy coasts and temporary ponds to mountain woods or rocky cliffs) (Table 1).

In the first phase, each partner collected the genetic material (mainly seeds) to be used for the production of new plants, while at the same time, each partner guaranteed the availability of the same material for future recovery or restoration programs through seedlots storage in seedbank and by implementing an "active collection" to be used for producing plants; in some cases, when it was not possible to make new seed collections, the seeds previously collected and stored in seedbanks were used. As an additional precautionary measure, accessions have been duplicated in the seedbanks of other project partners or, if necessary and appropriate, in other public institutions.

The plants were multiplied in botanical gardens and/or in public and private nurseries, depending on local needs, and the total number of the outplants produced varied depending on the type of plant (e.g. life form, seed availability, etc.), the presence/absence of biological or ecological limitations (e.g. difficulty to multiply the species, high mortality of the seedlings, etc.) or the translocation relevance (in term of number of plants; see Table 2).

Most of the translocations were planned as reinforcement for existing and threatened populations (44%), while other translocations consisted of reintroductions at sites where the plant had recently disappeared (not due to natural causes; 16%) or new populations at sites with no records of species occurrence in the past but with suitable ecological conditions (40%). In the majority of the latter two cases, selection of the microsites for planting was based on an expert-based criterion (Table 2). Our experience suggests that, although Species Distribution Modeling (SDMs) can efficiently delimit a range of potentially suitable areas, the final site selection should be made only by an expert whose decision-making takes into account many other variables, such as the real area availability, local political decisions, the acceptance by local communities, and so on.

Translocations have been implemented by using different plant material, mainly juvenile plants (64% of the total cases) and seeds (40%), followed by reproductive plants, seedlings and, in a few cases, bulbs (24, 14 and 2%, respectively); often a combination of different material type was used (40% of the total; Table 2). The effective realization of the translocations involved a wide range of protocols/options related to the target plants and the selected site for the translocation; in almost half of the cases, it was necessary to carry out preparatory actions (e.g. passive defense measures or control of invasive alien plants). Similarly, after outplanting, several actions were necessary (some of these were also repeated several times, such as supplementary watering in summer, control of alien species, etc.).

Finally, for each translocation a species-specific monitoring protocol was planned and implemented in order to ensure its sustainability; in particular, the monitoring activities were planned and implemented on a monthly basis for the 75% of the total translocations.

4. General lessons learned

All translocations were carried out recently (2017-2018), and for this reason, only preliminary data related to the first year after planting are available. It is therefore not possible to judge how successful particular programs and protocols were. However, some general lessons arise from the Care-Mediflora experience. It is widely recognized that translocations are useful tools to prevent the extinction of threatened species and to improve their conservation status (e.g. Maschinski and Duquesnel, 2006; Godefroid et al., 2011; Menges et al., 2016; Volis, 2017). However, CareMediflora approach confirms that severe limitations remain in the implementation of these conservation actions. In general, translocations are considered time-consuming activities, as they require several in-depth preparatory studies, a constant commitment to multiply the outplants, realize and monitor the activities, and a long-term monitoring plan to verify their effectiveness; in addition, translocation programs are considered economically expensive activities, because the pre- and post-translocation management actions required are generally high-priced. Actually, the factors that made translocation challenging, as well as stimulating, included limited human resources and availability of optimal hosting sites, bureaucratic difficulties encountered by working on both private and public properties, and the high uncertainty of success due to stochastic events.

Economic evaluation is relevant when working in territories particularly rich in threatened plant species, such as the Mediterranean islands, and when the economic resources are limited (Cogoni et al., 2013; Fenu et al., 2016). Although it is very complicated to make economic comparisons among different countries, the Care-Mediflora experience highlights that economic limitations are likely one of the most important constraints, and a detailed cost analysis is fundamental when planning a translocation. Considering all translocations carried out during the project, 80% of them required management actions that were complementary to outplanting, such as fencing, eradication of invasive alien species or weeding; these management actions are necessary both before and after translocation. Animal exclusion was the most frequent action, although this is not surprising given high grazing intensity throughout the Mediterranean. Even in other regions this is a common management practice when restoring rare and threatened plant species (Guerrant, 2012), greatly improving the demography of the introduced population over the short term (Godefroid et al., 2011; Fenu et al., 2016). In some cases, it was necessary to repeat or continue the post-translocation management actions such as eradication of alien species or control of the growth of surrounding vegetation. Clearly, the overall cost of a project will increase with the number and frequency of the after-care actions.

The costs of outplant production, on the other hand, are rather small compared to the total cost of a project, if efficient plant production protocols are available. Only in a few cases was outplant production expensive due to the involvement of specific treatments and fully controlled conditions during seedlings' growth in a greenhouse, as was the case with some relict plants represented by a single remnant population with limited sexual reproduction (e.g. Ribes sardoum Martelli, Urtica rupestris Guss., etc.). Thus, development of an efficient outplant production protocol, if the latter does not exist, is a way to reduce the cost of a translocation project. Significant cost reductions can also be achieved for the translocation program by the inclusion of researchers, public authorities, volunteers, and local stakeholders.

The Care-Mediflora experience also highlights the relevance of unexpected natural stochastic events that may affect the success of a translocation. In our project, such events mainly related to extreme weather conditions or the presence of feral animals, especially where no fences had been planned, which occurred in 40% of cases (Table 2). Extreme weather events are difficult to predict, but potential animal disturbance can be expected in the majority of the cases and, for this reason, fences (especially in the first years after transplanting) and/or protective fine chain-link fences should be used whenever possible.

Our experience confirms that, although there is a large number of endangered plants growing on private lands (and outside of the protected areas), translocations were more feasible on legally protected sites managed by public administration (76% of the total) rather than on private land (12% of the total; Table 2), as previously reported (e.g. Godefroid et al., 2011; Fenu et al., 2016); this should have clear implications for the choice of the optimal site, suitable habitat, and then the most suitable microsites.

In general, reinforcement of existing populations is considered a preferred option, while reintroductions and the creation of new populations are usually associated with greater uncertainty in terms of feasibility and success. To reduce this uncertainty, we used a precautionary approach in reintroduction projects by choosing only the sites where local extinction was due to human-mediated factors and not due to natural causes. Regarding the creation of new populations outside the known species range, several CareMediflora actions adopted a more positive perspective to assisted migration (colonization) than previous experiments. This perspective aligned with previous work that considered "facilitating species redistribution" (Bonebrake et al., 2018) as the only means to mitigate climate change-induced shifts in species range and to halt biodiversity loss (e.g. Hewitt et al., 2011; Thomas, 2011; Volis, 2016a, 2019).

From a practical point of view, two preliminary indications emerge after a year of monitoring. Generally, the initial survival rate seems related to the biological form: in the very short term, woody and shrub plant species reach the best performance with average survival rates of ca. 50% and therefore they seem to better tolerate the transplant shock. Secondly, after one year, the best results on average were obtained for translocations carried out using juvenile and reproductive plants; although exceptions have been observed, this preliminary indication is consistent with previous studies that have highlighted the advantages of using juvenile and reproductive plants rather than seeds or seedlings because they have usually higher survival rates than seedlings or seeds (Godefroid et al., 2011; Albrecht and Maschinski, 2012; Liu et al., 2015; Menges et al., 2016). More specifically, although it is not reasonable to draw conclusions after a few months of observation, the preliminary data indicate a lack of significant differences between translocations carried out using/planting juvenile or reproductive plants. This finding, if confirmed in the future, suggests that the use of juvenile plants would be preferable because they reduce the costs of production/ maintenance/cultivation of plants in nursery, specifically, as in our case, when working with forbs, geophytes and (half)shrubs that reach maturity within a few years and that the use of plants older than one-year old has no advantage. On the other hand, it is difficult to assess success in projects performed on annual plants or with direct sowing of seeds. The success of such projects can rarely be evaluated in the short-term, because seeds may enter the soil seed bank and persist there for years. As noted by Albrecht et al. (2018), little is understood about translocating annuals, for which there is no alternative to sowing. This, however, requires long-term monitoring. Additionally, the true success of translocation - regardless of life history - is only evident in successful recruitment of new generations (e.g. Bell et al., 2003; Godefroid et al., 2011; Fenu et al., 2016). Prolonged monitoring increases the probability that rare events with population effects will be detected (e.g. Monks et al., 2012; Duquesnel et al., 2017). In addition, when results of a translocation are evaluated through sufficiently long monitoring, poorly performing methods can be discarded and replaced with more effective ones (Kaye, 2008). The initial monitoring frequency during the Care-Mediflora project (monthly in majority of the cases, 75%), was found to be too costly for the Mediterranean islands (as earlier suggested by Cogoni et al., 2013 and Fenu et al., 2016). For this reason and the reasons described above, each partner committed to a less intensive but long-term monitoring protocol including at least five years after completion of the project.

5. Concluding remarks

The Mediterranean insular territories, which share an extraordinary rate of endemism coupled with a remarkable degree of environmental and human-related threats, provide an opportunity to combine different methods and methodologies within a common conservation strategy focusing on endangered plants. In addition, due to their limited area, discrete nature, and simplified food webs, islands can represent "natural laboratories" for ecological studies, including translocation experiments. The CareMediflora project represents the first attempt to develop common strategies and an opportunity to join and integrate methods and methodologies focused on threatened plant conservation in the Mediterranean islands. The project's actions represent a step forward for the conservation of the Mediterranean flora, and perhaps a basis for planning conservation measures for the other species threatened with extinction if our experiences are replicated in partner countries (at a larger scale), as well as in other Mediterranean countries with similar environmental conditions. Furthermore, translocations have a relevant social-cultural impact and strengthen fruitful collaborations among national and regional administrations, as well as NGOs and local stakeholders (Maschinski and Duquesnel, 2006). For this reason, an important activity of Care-Mediflora project was dedicated to sharing knowledge and experiences among partners and adopting common protocols. Care-Mediflora project participants from all islands shared differences in translocation protocols at meetings to develop the technical aspects, refine methodologies, and plan successful in situ conservation actions. At the same time, each partner has actively involved the local and regional authorities and local stakeholders (in particular in the monitoring activities) to make the translocations more effective.

Finally, the translocations carried out during the Care-Mediflora project represent an important contribution to the achievement of the EU Biodiversity Strategy to 2020 (i.e. target 6) and several Aichi Targets (e.g. 11, 12 and 19) but in particular to the implementation of the in situ conservation measures advocated by Aichi Target 12.

None declared.

MAVA Foundation is gratefully acknowledged for financially supporting (80%) the CARE-MEDIFLORA project. The authors wish to thank all those people who gave their contribution to the CareMediflora project. The authors are also grateful to the Editor and the anonymous reviewer for helpful comments.

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2019.03.001.