SCIENCE AT WORK FOCUS LAB WATCH INNOVATION © U.S. GEOLOGICAL SURVEY 1996/ COMMONS WIKIMEDIA A Ground-Breaking Theory Vital for trapping carbon deep underground instead of clogging up the atmosphere, tectonic plates— whose shifts can cause earthquakes and volcanic activity—make up the lithosphere, the planet’s outermost layer. But how did this once uniform layer end up as separate plates? Using mathematical models to trace the behavior of tiny mineral grains in lithospheric rock, geophysicists Yanick Ricard of the LGLTPE1 together with David Bercovici from Yale University now offer a ground-breaking explanation for the birth of tectonic plates.2 Rock evidence shows that 4 billion years ago, when the Earth’s shell was still intact, convection movement between the cool surface and the hot mantle first drew single pieces of crust into the mantle, deforming the lithosphere. It would take another billion years for the shell to finally split into plates free to slide under one another and dip into the mantle in the process of subduction, at the core of tectonic activity. So what factors at work during the billion-year lag ultimately detached the plates? Studies on rock deformation mechanisms alone offered no answer. “Our intuition,” explains Ricard, “was that on these timescales, mineral-grain size evolves.” The team therefore decided Geophysics. New research reveals how our planet’s crust split up into separate plates. BY FUI LEE LUK 1. Laboratoire de géologie de Lyon, terre, planètes, environnement (CNRS / Université Lyon-I / ENS de Lyon). 2. D. Bercovici and Y. Ricard, “Plate tectonics, damage and inheritance,” Nature, 2014. 508: 513-6. 8 CNRS INTERNATIONAL MAGAZINE to create an evolution model that took into account both time and deformation. More specifically, they looked at two rock components with similar properties to the two mineral types (olivine and pyroxene) found in peridotite, the main rock in lithospheric plates. What their model reveals is a crucial self-weakening cycle: mineralgrain size tends to shrink in rock deformed by mantle convection. As tinier grains are more malleable, deformation and grain-shrinkage recurred in the arising fragile zone until breakage point was reached. Fragile zones can potentially “heal” over time via the regrowth of grain size, but on Earth, the interaction of two types of minerals blocked the grain coarsening. As Ricard points out, “large olivine crystals have trouble forming from smaller ones if they are mixed with pyroxene, since grains of the same type are less likely to meet.” Moreover, due to the Earth’s moderate temperatures, the healing process, which requires high temperatures, was slower than the weakening process. In contrast, on Venus, also modeled by the team, the 500°C temperatures promoted “much faster healing,” which explains why this planet— otherwise similar to Earth in size, composition, and gravity—lacks tectonic plates. According to the team, the Earth’s weak zones accumulated and joined up, eventually forming a network of boundaries that individualized the tectonic plates. The model suggests that this process took approximately one billion years, which fits perfectly with the observed data. ii LAB WATCH Neuroscience This map shows the location of the 15 largest tectonic plates of our planet. Broken Sleep Helps Remember Dreams BY FUI LEE LUK Why do some people remember dreams more easily than others? According to a CRNL1 team, it all comes down to activity levels in two forebrain areas. Observing volunteers via Positron Emission Tomography (PET) scans, which provide 3D imaging of body functions, the neuroscientists not only offer the first empirical proof of the forebrain’s key role in dreaming, but also identify factors that help good dream recall—chiefly broken sleep.2 To check an untested theory that forebrain mechanisms steer dreaming, the team studied 41 volunteers separated into 2 types of dreamers: 21 “high recallers”—who remember five dreams a week on average—and 20 “low recallers”—who only recall dreams about twice a month. Sleep-deprived for 36 hours, the subjects were then wired to PET scanners to measure their spontaneous brain activity, and finally allowed to sleep. A true feat, team leader Perrine Ruby laughs, for the subjects “had to lie perfectly still and flat, without pillows, and with catheters in their arms.” Confirming the forebrain’s role in dreaming, results showed that high recallers have stronger spontaneous brain activity in two of its zones. Firstly, the medial prefrontal cortex, whose contribution to mental images and associations may make dreams more story-like, thus easier to recall. Secondly and notably, the temporo-parietal junction. Greater activity in this zone, which is involved in drawing attention to outside stimuli, backs the team’s earlier findings that high recallers are more reactive to sounds and wake up twice as often during sleep than low recallers.3 Broken sleep would in fact enhance dream recall. “A sleeping brain can’t memorize new information,” explains Ruby, therefore brief waking moments allow “dreams to be encoded into memory. ” The team’s next objective is to repeat the test “over an undisturbed night’s sleep.” ii © INSERM Frequent dreamers show heightened activity in the brains’ temporoparietal junction. 1. Centre de recherche en neurosciences de Lyon (INSERM / CNRS / Université Lyon-I). 2. J.-B. Eichenlaub et al., “Resting brain activity varies with dream recall frequency between subjects,” Neuropsychopharmacology, 2014. 39: 1594-1602. doi:10.1038/npp.2014.6. 3. J.-B. Eichenlaub et al., “Brain reactivity differentiates subjects with high and low dream recall frequencies during both sleep and wakefulness,” Cerebral Cortex, 2014. doi: 10.1093/cercor/bhs388. 9 SUMMER 2014 N° 34 SCIENCE AT WORK © TAKASU/FOTOLIA Cryptography Legal Hacking BY BRETT KRAABEL CNRS researchers recently solved a 40-year-old puzzle by decoding one variant of the discrete logarithm problem, a formula that is used to secure objects like RFID tags.1 Why would researchers spend public money trying to hack security algorithms? To show that cryptographic methods based on these algorithms are vulnerable to modern computing attacks, and warn authorities not to use them in security applications. All security algorithms are based on formulas that take a clear (i.e., uncoded) symbol as input and, by using a number called a “key,” return a coded symbol as output. For a given input, different keys will produce different outputs and, without knowing which key is used, it is virtually impossible to determine the original message from the encoded one. A good algorithm is one whose security can be improved simply by increasing the number of bits in the secret key. For instance, the key used for credit-card security algorithms is regularly lengthened to combat the ever-increasing computing power of hackers. In 2009, this key length was doubled, meaning that instead of one year, it would now take a million years to crack the code. If this change in key length only resulted in doubling the time it would take to crack the code, then the security algorithm could not stay ahead of computing power. Thus, when attacking a cryptographic method, explains Pierrick Gaudry from the Loria,2 “the Holy Grail is to find an attack algorithm whose computation time increases [almost in proportion to] the number of bits in the key.” For a variant of the discrete logarithm problem, Gaudry and colleagues did just that: they devised an attack algorithm whose calculation time increases almost in proportion to the number of bits in the key. This has led authorities to withdraw their endorsement for using this discrete logarithm variant in cryptography and has provided a new algorithm with which to test other cryptographic methods. “We protect citizens by guaranteeing the security of cryptographic applications,” concludes Gaudry. ii Authorities constantly have to lengthen credit card security keys to prevent fraud. 1. R. Barbulescu et al., “A Heuristic Quasi-Polynomial Algorithm for Discrete Logarithm in Finite Fields of Small Characteristic,” Advances in Cryptology – EUROCRYPT 2014, Lecture Notes in Computer Science, 2014. 8441: 1-16. 2. Laboratoire lorrain de recherche en informatique et ses applications (CNRS / Université de Lorraine / Inria). Biology Fighting Malaria Relapse BY EMMA WALTON Many people who have had malaria can relapse years after the initial infection. Such relapses are due to a dormant form of the parasite, so far poorly accessible to researchers. Now, a team led by Georges Snounou1 and Dominique Mazier2 has succeeded in cultivating it for the first time.3 Malaria is caused by Plasmodium, a single-cell parasite carried by the mosquito Anopheles. Following a bite from an infected mosquito, tiny parasitic sporozoites enter the bloodstream and pass into the liver. There, they multiply to form thousands of merozoites that enter the red blood cells and cause the symptoms of malaria. Two species of Plasmodium, P. vivax and P. ovale, also produce dormant forms in the liver. Named “hypnozoites,” they can wake up without warning and induce relapses. So far, the study of hypnozoites has involved infected monkeys and “the occasional, gallant volunteer,” says Snounou. But as he points out, the ratio of hypnozoites to liver cells is so low that studying them is like looking for a needle in a haystack. An ideal model would consist of infecting liver cells in a dish. Yet “liver cells fail to grow in culture,” explains Mazier, and infection with Plasmodium causes many cells to detach and die. Now, by adding liver cancer cells to plug the holes left by dying cells and providing a soft support, Mazier’s team has managed to grow infected cells in culture for up to 40 days, approximately four times longer than previous attempts. The model was put to the test by screening for anti-hypnozoite compounds. Interestingly, one molecule awakened some of the hypnozoites. This led to a therapeutic strategy called “wake and kill,” involving one drug to awaken the hypnozoites, and another to kill them. “This may be the key to a radical cure,” concludes Mazier. © J.-F. FRANETICH ET L. DEMBÉLÉ/CIMI-PARIS In liver cells, the parasite P. cynomolgi can form thousands of merozoites ready to infect blood cells (large green shape), or stay in a dormant form (small green dot). 1. Centre d’immunologie et des maladies infectieuses (INSERM / Université Pierre et Marie Curie / CNRS). 2. Service de parasitologie-mycologie (AP-HP / Hôpitaux Universitaires Pitié-Salpêtrière). 3. L. Dembélé et al., “Persistence and activation of malaria hypnozoites in long-term primary hepatocyte cultures,” Nat. Medicine, 2014. doi: 10.1038/nm.3461. 10 CNRS INTERNATIONAL MAGAZINE LAB WATCH Tough as Pearls Materials. A new and stronger type of ceramic inspired by naturally-occurring microstructures could soon find industrial applications. BY ARBY GHARIBIAN © S. DEVILLE, F. BOUVILLE/LSFC Judging by their frequent appearance in archeological digs, ceramics were as useful to humans thousands of years ago as they are today. Small and lightweight ceramic parts could find many applications from car engines to large industrial furnaces. Yet despite important advances in materials science, these objects still have the significant drawback of being fragile, limiting their many potential industrial uses. A new technique, developed by French laboratories led by a CNRS team,1 has now produced ceramics up to ten times tougher than those manufactured using traditional processes.2 “We looked no further than nature,” explains team leader Sylvain Deville. “More specifically, we studied abalones, small marine mollusks with surprising properties. Despite the fact that 95% of their mother-of-pearl shells is made up of a fragile carbonate mineral known as aragonite, they are able to withstand great stress, much more than their composition would suggest.” The scientists surmised that microstructure played an important role in the shell’s strength and resilience. A closer look revealed that the aragonite consisted of small platelets organized into highly-textured layers piled on top of one another. This last detail is key, as it forces cracks to move along a curving and twisted path, thus lessening their scope and limiting their propagation. The challenge now was how to artificially organize the materials traditionally used for ceramics into this microstructure. Researchers suspended a common ceramic powder— called alumina—in water, and then used freeze-casting, a gradual process that lets the user control the speed and direction of ice crystal formation. They then dried the ice crystals, making it possible to press the platelets into the sinuous alignment that increases toughness in the abalone shells, before densifying the remaining substance at high temperature to yield the final material for ceramic production. “Our technique produces tougher ceramics without the need for amalgamating metal or polymers, which can compromise functionality and stability at high temperatures,” adds Deville. “From an industrial perspective, this method seems compatible with large-scale industrialization, and should add little to production costs for techniques already in use.” ii 5 µm Natural (left) and synthetic (right) structure of mother-ofpearl, showing the highlytextured layers. 1. Laboratoire de synthèse et fonctionnalisation des céramiques (CNRS / Saint-Gobain). 2. Florian Bouville et al., “Strong, tough and stiff bioinspired ceramics from brittle constituents,” Nature Materials, 2014. 13: 501-14. 11 SUMMER 2014 N° 34 SCIENCE AT WORK © ESA / COLLABORATION PLANCK The magnetic field of the Milky Way as seen by Planck (darker areas show strength, ridges show direction). Our Galaxy’s Magnetic Footprint BY TOM RIDGWAY Astronomy. The Planck mission has released a new map of our galaxy’s magnetic field, using measures of the polarized light emitted by interstellar dust. Our galaxy’s magnetic field helps slow gravitational collapse in regions where stars form. Knowing its structure is therefore necessary to understand star formation,” explains Jean-Philippe Bernard, researcher at the IRAP1 and member of the Planck collaboration, the space observatory operated by the European Space Agency (ESA), which provided the data. To create a map of this field, Planck’s detectors measured the light emitted by polarized dust particles in the interstellar medium. “This light has a slight linear polarization, at right angles to the direction of the magnetic field,” says Bernard. This means that by following polarized light, researchers are able to map the direction of the magnetic field and so reveal its overall structure. Along the Milky Way, the field is largely parallel to the galactic plane, yet “it shows convoluted structures at higher latitudes, unveiling the complexity of the field geometry,” adds Bernard. This research is the latest step in fulfilling one of Planck’s main missions: mapping the brightness of the universe’s cosmic microwave background (CMB), the relic radiation of the Big Bang. “These results will give us more information on how to subtract our galaxy’s polarized signal from the cosmological data,” says Bernard. “This is essential for precisely measuring the polarization of the CMB, in particular revealing the much-anticipated presence of primordial gravitational waves.” Generated in the first instants after the Big Bang, these waves could give researchers invaluable insights into the birth of our universe and the formation of the first stars several hundred million years later. The analysis of light polarization was performed using data provided by Planck before it was shut down in October 2013, and was the subject of four articles recently submitted to Astronomy & Astrophysics by the Planck collaboration, including CNRS researchers. ii 1. Institut de recherche en astrophysique et planétologie (CNRS / Université Toulouse-III). 12 CNRS INTERNATIONAL MAGAZINE LAB WATCH To be followed... BY TOM RIDGWAY PLATO: A SPACE-BASED TELESCOPE TO SEARCH FOR EARTH-LIKE PLANETS The European Space Agency’s (ESA) Science Programme Committee has selected the six-year PLATO (Planetary Transits and Oscillations of Stars) as its third M-class mission. Scheduled to launch in 2024, the satellite will use 34 separate small telescopes and cameras to discover and characterize Earth-sized planets and super-Earths in their own stars’ habitable zones. This data will then be combined with ground-based radial velocity observations to calculate planetary mass and radius, and infer density. The satellite’s construction, launch, and flight operations will be run by the ESA, while a consortium of European laboratories (including a number of CNRS-affiliated institutes in France) will collect and analyze the data. The PLATO mission will identify and categorize thousands of exoplanetary systems. © C. CARREAU / ESA The 3D spectrograph MUSE, at the VLT in Chile, will observe distant galaxies. FIRST LIGHT ON MUSE The European Southern Observatory’s Very Large Telescope in Paranal (Chile) recently welcomed the seven-ton Multi-Unit Spectroscopic Explorer (MUSE). The instrument, developed over a decade by a pan-European consortium headed by the CRAL1 in France, will use 24 wide field-of-view spectrographs to allow researchers to create 3D views of the universe. Described as a “fantastic time machine” by the project’s principal investigator Roland Bacon, the instrument will look into the earliest moments of our universe and the mechanisms of galaxy formation, as well as the motion of matter and the chemical properties of nearby galaxies. © E. LE ROUX/CNRS PHOTOTHÈQUE/UNIVERSITÉ CLAUDE BERNARD LYON 1/ESO A photo gallery on MUSE is available online Artist’s rendering of the approved CTA observatory. NEGOTIATIONS FOR THE CTA SITE BEGIN The Cherenkov Telescope Array moved one step closer to reality last April. Delegates representing 1000 scientists from 170 research institutes in 28 countries voted to start negotiations for two potential base sites in the Southern Hemisphere. Once built in either Aar (Namibia) or at the European Southern Observatory (ESO) (Chile), the CTA will comprise more than 100 Cherenkov 23-, 12-, and 4-meter telescopes. These will provide an order-ofmagnitude jump in the study of cosmic rays and their role in the universe, as well as shed light on the nature and variety of particle acceleration around black holes. A decision regarding a final site will be made at the end of the year. © COLLABORATION CTA 1. Centre de recherche astrophysique de Lyon. 13 SUMMER 2014 N° 34 SCIENCE AT WORK NEW HINTS ON OBESITY Obesity, a condition that kills 2.8 million people worldwide each year, has recently been the subject of important findings. Focusing on salivary amylase (AMY1), a human protein responsible for digesting complex sugars (starches), an international team led by Philippe Froguel1 identified a region in chromosome 1 containing the gene encoding AMY1, whose number of copies varies from 1 to 20 depending on the individual.2 They found that having a low copy number of AMY1—and hence low amylase levels in the blood—favors obesity and that the risk of becoming obese increases by 20% with each unit drop in the number of copies. Another study3 led by Serge Luquet4 demonstrated that long-lasting exposure of the brain to triglycerides in mice increased their need for food rewards and craving for rich foods, while reducing their locomotor activity. Triglycerides in food may therefore act as hard drugs on the brain’s reward system. GLUING ORGANIC TISSUE Last year, Ludwik Leibler and his team5 used nanoparticles to glue gels and biological tissues together in vitro.6 Now, the same team, working in collaboration with that of Didier Letourneur7 has shown that this method can also be used in vivo, in mice, to close deep wounds in seconds, providing aesthetic healing.8 The technique has also proved effective to repair soft-tissue organs, or to attach a membrane to a beating heart for cardiac cell therapy. © S. DE GRISSAC/CNRS/CEBC Capture of a south polar skua (Stercorarius maccormicki ) in Adélie Land (Antarctica). Environment Mercury’s Long-term Effect on Skuas BY EDDY DELCHER The long-term fecundity of skuas, large predatory seabirds, is being jeopardized by mercury contamination, according to a recent study1 by researchers from the CEBC2 and from the LIENSs3 laboratory. The research team highlighted a direct correlation between concentrations of the highly-toxic metal found in skuas’ blood and their reproduction rate by monitoring two different skua species, in Adélie Land and the Kerguelen Islands, over a 10-year period. “The impact of mercury contamination on animals was already known, but had never been assessed in the long term,” says Olivier Chastel from the CEBC, who led the study. To analyze each individual’s mercury levels, blood samples were regularly taken from a hundred tagged skuas, while the researchers also monitored nesting sites to assess breeding rates. “We studied the skua because it sits high on the food chain, and is therefore more affected because it accumulates the mercury contained in other animals,” explains Chastel. The team predictably found that mercury contamination lowers birth rates, as it disrupts the secretion of hormones vital to the reproduction process. A more unexpected occurrence is that the Adélie Land skuas, which display the lowest mercury levels, are the most affected by the contamination. “This can be explained by the increasing presence of other pollutants in that area, which, coupled with the region’s harsh climate, may worsen the effects of mercury poisoning,” says Chastel. The team is now studying more Antarctic species to assess the impact of other heavy metals and pollutants such as PCB and pesticides. It recently published an article describing the consequences of this pollution on the wandering albatross4. ii 1. A. Goutte et al., “Demographic responses to mercury exposure in two closely-related Antarctic top predators,” Ecology, 2014. DOI: 10.1890/13-1229.1. Funded by the Agence Nationale de la Recherche and the Institut polaire français Paul Émile Victor. 2. Centre d’études biologiques de Chizé (CNRS / Université de La Rochelle). 3. Littoral, environnement et sociétés (CNRS / Université de La Rochelle). 4. A Goutte et al., “Demographic consequences of heavy metals and persistent organic pollutants in a vulnerable long-lived bird, the wandering albatross,” Proc. R. Soc. B, 2014. doi:10.1098/rspb.2013.3313. 14 CNRS INTERNATIONAL MAGAZINE WORLDWIDE PARTNERSHIP ON LOCATION NEWSWIRE CNRS Joint Units: the North American Connection © L. FRÉCHETTE / UMI-LN2, 3IT - UNIVERSITÉ DE SHERBROOKE © S. ECOFFEY / UMI-LN2/ 3IT-UNIV. DE SHERBROOKE / ST MICROELECTRONICS Low-power consumption CMOS electronic chip developed at the LN2. A microturbine for energy recovery developed at the LN2. Joint research. The first meeting bringing together the CNRS North American International Joint Units (UMIs) was held at the French embassy in Washington in May. While their research fields and organization differ, these UMIs are equally successful. BY ISABELLE TRATNER The CNRS initiative, 20 years ago, to set up a North American research area is finally a reality,” said Xavier Morise, director of the CNRS office in Washington, at the inaugural meeting of the organization’s International Joint Units (UMI) (see box) in the US capital last May. In 2014, no fewer than 12 of the 35 CNRS UMIs worldwide are located in North America, including 7 in the US, 4 in Canada, and 1 in Mexico. Not to mention the UMIFRE CEMCA1 in Mexico, a social science research unit jointly run by the CNRS and the French Ministry of Foreign Affairs. Structuring collaboration A UMI is often the result of a longstanding collaboration between researchers. This is the case of the CIRHUS2 UMI in archeology, founded in 2008 and based at New York University. CIRHUS director Randall White has collaborated for more than 20 years with the French researchers who are now partners in this UMI. Similarly, the LN23 UMI was created in 2012 for the study of 3D nanoelectronics, biosensors, and energy after nearly 15 years of teamwork between researchers from the INSA in Lyon (France) and the University of Sherbrooke (Canada). A magnet for industry Setting up a UMI not only confers the advantage of shared facilities and, consequently, the possibility of much larger investments, but also offers greater access to industrial partners. “It is a win-win situation for all involved,” says Abdelkader Souifi, co-director of the LN2 UMI, who has notably established a partnership between the LN2 and STMicroelectronics for the 2012-2017 period. “For Canada, which does not yet manufacture CMOS chips, this cooperation is crucial,” adds Souifi. Ditto for the COMPASS4 UMI, founded in 2009, whose focuses include energy transfer and storage. Solvay, its 1. Centre d’études mexicaines et centraméricaines. 2. Center for International Research in the Humanities and Social Sciences. 3. Laboratoire nanotechnologies et nanosystèmes. 4. Complex Assembly of Soft Matter. 5. Pacific Institute for the Mathematical Sciences. 6. Laboratoire franco-mexicain d’Informatique et automatique. 7. Centre de recherches mathématiques. 8. Institut des hautes études scientifiques. 36 CNRS INTERNATIONAL MAGAZINE industrial partner, finances one to three students each year and several permanent researchers, and also provides the working premises. In return, the company benefits from the expertise of the UMI members and facilities of the University of Pennsylvania. Added value for students Industrial partnerships and funding, although important, are not necessarily the primary goal. Several UMI directors have set education as a priority—“from undergraduate to post-doctoral level,” specifies Alejandro Adem, director of PIMS-Europe,5 the first UMI in mathematics set up in Canada in 2007. The results speak for themselves, as Rogelio Lozano, director of the Mexico-based LAFMIA6 UMI, specialized in robotics, points out. “One of the main benefits of the UMIs is that they make students more attractive to potential employers; many of those who have gone through the LAFMIA have since been recruited by the private sector,” he adds. Across the border in southern California, Guy Bertrand, director of a UMI in catalytic chemistry at the University of San Diego, makes the same observation. “Most of our students are hired after studying at the UMI, many in France, and some at the CNRS.” Mirror sites The Georgia Tech-CNRS UMI, which focuses on secure networks and smart materials and housing, is uniquely positioned in this research landscape: the first UMI to be created in France, it was inaugurated in Metz in 2006 before going on to establish a “mirror” site at the Georgia Institute of Technology, its partner in Atlanta (US). “Creating mirror UMIs would help attract foreign researchers to France,” says Marcel Babin, director of Takuvik, a UMI created in 2011 at the University of Laval, Quebec (Canada), to study the impact of climate change on the Arctic. “It would also help pursue the research and preserve the skills acquired by the French researchers who have worked at the UMI.” The CRM7 UMI, also in Canada, which covers all areas of mathematics, is currently in talks to create a mirror site at the IHÉS,8 a center of excellence in mathematics for the Paris region. A much-envied structure Funding from industry, abundant research output, international networking, exchanges of researchers, students, and post-docs—UMIs offer many advantages, making them the envy of the CNRS’s foreign partners. Babin was even invited by a German colleague to meet with Chancellor Angela Merkel, to try and persuade her to set up equivalent structures for Germany. In his closing address, attended by representatives from the Canadian and Mexican embassies, François Delattre, the French ambassador to Washington, stated that science, research, and innovation are high priorities for France. No doubt UMIs are an excellent means of pursuing this goal. ii PARTNERSHIP Loading a semiconductor wafer in the growth chamber of an epitaxial reactor in vapor phase at the Georgia Tech UMI. © P. COUPEL © A. CHEZIERE / CNRS PHOTOTHÈQUE The Takuvik UMI uses a CTD/ rosette sampler to analyze the water column. What is a UMI? International Joint Units (UMI) enable the CNRS to work with one or more foreign partners— and occasionally industrial ones—on joint research programs in a dedicated facility. The various partners assign researchers to the UMI for a given period, from a few months to several years. The UMI’s legal status makes it eligible for funding from agencies and foundations in the partner countries as well as from the European community. Significant private funds may also be available when industrial partners are involved. 37 SUMMER 2014 N° 34