Search

Improving the nutritional and health properties of food products, e.g. infant milk formula, by the addition of functional ingredients is of primary importance. This study focused on bioactive milk polar lipids (PLs) recovered from dietary sources that are of increasing interest. The chemical compositions of buttermilks and butter serums were determined and the modulation of the fatty acid composition of milk PLs was investigated. Butter serums contain a higher amount of milk PLs than buttermilks (88 vs. 13-18g/kg dry matter), with a higher proportion of sphingomyelin (34 vs. 19% of PLs, respectively) interestingly close to human milk PL profile. Butter serums are also interesting sources of choline, an important nutrient for infant brain development. We demonstrated that the unsaturated fatty acid content of milk PLs recovered in the buttermilks and the butter serums, mainly the amount of C18:3n-3 and C22:6n-3 (DHA) that are of nutritional interest, can be increased by dietary strategies. This work opens perspectives for a better valorization of milk PLs in human nutrition (both infants and adults) to benefit their functional, nutritional and health properties.

Ce poster présente une action en cours financée par le projet I-Site de Clermont-Ferrand (http://cap2025.fr/le-projet-i-site/), en particulier le Challenge 1 sur les agro-écosystèmes durables dans un contexte de changement global. Cette action vise à étudier le potentiel fourrager d'espèces ligneuses potentiellement intéressantes en agrosylvopastolralisme pour l’environnement auvergnat en 2050. Afin d'évaluer ce potentiel, un phénotypage multi-traits a été initié pour le choix d’espèces ligneuses, en se basant sur un croisement entre valeur alimentaire des feuilles et des caractéristiques écophysiologiques déterminant la survie des espèces d'intérêt (i.e. résistances à la sécheresse et au gel). Les résultats préliminaires issus d'une revue bibliographique, et de mesures effectuées en 2019 (résistance à la sécheresse et résistance au gel, valeurs alimentaires in vitro: matière sèche, matière organique, matière azotée total, parois) sont présentés pour un panel d'espèces ligneuses présentes dans l'étage moyen-montagnard autour de Clermont-Ferrand.

The antimethanogenic compound 3-nitrooxypropanol (3NOP) has been proven efficient to reduce methane (CH4) emissions from ruminants but it has been tested mainly on high-producing dairy cows fed high-concentrate diets. The aim of this work was to test the effect of 3NOP on CH4 emissions from medium-yielding cows fed a high-forage diet. Twenty-eight lactating dairy cows were recruited within 7-11 days after calving and distributed in 2 balanced groups in a randomized block design. The treated group received 3NOP (60 mg/kg DM basis) in a total mixed ration during 105 days and the control group received a placebo. The diet had 75% forage (52% corn silage, 17% hay, 6% straw) plus 25% concentrate and was distributed once a day after the morning milking. Individual daily kinetics of CH4 emissions were quantified using the GreenFeed system (C-Lock Inc., Rapid City, SD, USA). Intake and milk production were recorded daily, and milk composition (fat, protein, lactose, urea) twice a week. Intake was reduced with 3NOP (-1.6 kg DM on average; P = 0.01) whereas milk production was similar between groups (34.7 kg/d of energy corrected milk on average; P = 0.28). Milk composition did not vary between groups. Feed conversion efficiency in animals tended to increase with 3NOP (1.37 vs. 1.46 kg milk/kg DM intake on average; P = 0.07). Methane emissions, CH4 yield and CH4 intensity were lower in the treated group throughout the 105 days (on average -31% g/d, -25% g/kg DM intake, and -30% g/kg energy corrected milk, P < 0.0001). When monitoring daily kinetics of CH4 emissions, we observed that the major mitigation effect of the 3NOP occurred after feeding by preventing the post-prandial peak of CH4 emissions (P < 0.001). In conclusion, the CH4-reduction effect persisted for the duration of the study in early lactation dairy cows fed a 75% forage-based diet. The additive consistently prevented the CH4 production peak after the main meal of the day (P < 0.001). In systems without a total mixed ration, 3NOP should still have a significant antimethanogenic effect if provided throughout the main daily feeding period.

In Press, Corrected Proof; Although the effects of cow diet on cheese sensory properties have been well documented, the putative interactions between the biochemical and microbial milk components and their respective roles in the development of the sensory properties of cheeses have yet to be explored in depth. The aim of this study was to evaluate the specific contribution of milk fat composition to the formation of cheese sensory properties. Two creams with different fat compositions were obtained from cows fed either pasture or maize silage. Cheeses were manufactured from the same skim milk (identical chemical and microbial composition) with either the pasture- or maize silage-origin pasteurized cream added. The gross composition and microbial composition of milks did not vary with cream origin. In milks and cheeses, the fatty acid (FA) profiles were modified by the origin of the cream. The concentrations of C18:0 and unsaturated FA such as cis-9 C18:1, trans-11 C18:1, C18:3n-3, total conjugated linoleic acids, and mono- and polyunsaturated FA were higher in milks and cheeses with the pasture-origin cream than in those with the maize-origin cream. In contrast, the maize milks and cheeses had higher concentrations of short- and medium-chain saturated FA, C16:0, and C18:2n-6. The level of lipolysis was 11% in the cheese rind and only 0.30% in the cheese core. The rind of pasture cheeses had a higher concentration of free C18:0 and C18:3n-3 and a lower concentration of free C14:0 and free C16:0 than the rind of maize cheeses. The levels of major microbial groups were similar in pasture and maize cheeses at different stages of ripening. The pasture cheeses had a more elastic and creamier texture, a yellower color, and a thinner rind than the maize cheeses, but the odor and aroma of cheeses were not affected by the origin of the cream, despite a few modifications in the balance of volatile compounds from FA catabolism. Based on these results, we conclude that milk fat composition modulated by cow diet had a direct role in the texture of the cheese but no effect on flavor. The high degree of lipolysis in cheese rind, along with the higher concentration of long-chain unsaturated free FA in pasture cheeses may be responsible for antimicrobial activity, which could explain differences in the appearance of cheese rind.

International audience; Modulating the assembly of the ruminal microbiota might have practical implications in production. We tested how an early-life dietary intervention in lambs influences the diversity and function of the ruminal microbiota during and after the intervention. Microbiota resilience during a repeated dietary intervention was also tested. The treatment, aiming to mitigate enteric methane emissions, combined garlic essential oil and linseed oil. Fifty-six lambs and their dams were allocated to two groups and treatment (T1) or placebo (C1) was drenched from birth until 10 weeks of life. Lambs were weaned at 8 weeks. From 16 to 20 weeks, lambs in each group were divided in two subgroups that received (T1–T2 and C1–T2) or not (T1–C2 and C1–C2) the same treatment. Measurements were done at 8, 14, and 20 weeks. Average daily gain was similar between groups. Methane production was reduced by treatment at 8 and 20 weeks but at 14 weeks it was similar between C1 and T1. Interestingly, early-life treated lambs displayed a numerical increase (P = 0.12) in methane emissions at 20 weeks compared with non-treated lambs. Concentration of VFA was not affected by the intervention at 8 or 14 weeks but a lower concentration was observed in T2 lambs compared with C2 at week 20. Metataxonomics (rRNA gene) revealed differences in archaeal communities between groups of lambs when treatment was applied (weeks 8 and 20); whereas, in accord with methane emissions, these differences disappeared when treatment was discontinued (week 14). Protozoal community structure was not affected by treatment. In contrast, bacterial community structure differed between treated and non-treated lambs during and after the intervention. Rumen and urine LC-MS and NMR metabolomics at week 20 separated C2 from T2 lambs and correlation analysis highlighted interactions between microbes and metabolites, notably that of methylated compounds and Methanomassiliicocceae methanogens. This study demonstrates that a long-term early-life intervention induced modifications in the composition of the rumen bacterial community that persisted after the intervention ceased with little or no effect on archaeal and protozoal communities. However, there was no persistency of the early-life intervention on methanogenesis indicating resilience for this function.

Les mesures de Production Laitière (PL) des vaches allaitantes consistent à évaluer la quantité de lait bu par le veau grâce à deux pesées répétées avant et après tétée, mais elles restent difficilement réalisables et limitées aux unités expérimentales. Une base de données expérimentales regroupant 12 000 mesures de lait bu, réalisées depuis 15 ans dans les troupeaux charolais, limousins et salers de l’Herbipôle (plus de 1 000 vaches contrôlées et 7 mesures au moins par vache et par lactation) a été constituée pour actualiser nos connaissances quantitatives sur la production de ces vaches et en étudier les principaux facteurs de variation. La race influence la production laitière (1 600 ± 313 kg lait/lactation en race limousine, 1840 ± 355 kg en race charolaise et 2250 ± 470 kg en race salers) et on observe toujours environ 11% d’écart entre primipares et multipares soit environ 1 kg/j de plus pour les multipares. Dans un système « classique » de production de broutards, avec vêlage d’hiver et conduite semi-extensive au pâturage, deux pics de production laitière sont observés. Le premier a lieu un mois après le vêlage et le second, plus important, peu après la mise à l’herbe. Ces pics traduisent une augmentation de production + 0,8 à + 2,3 kg selon la race, la parité et la période de vêlage. L’analyse de la base de données a permis de quantifier la persistance de la lactation et sa variabilité (diminution de la production de lait de 17 à 27 g de lait bu par jour). La persistance est meilleure pour les vaches faiblement productrices (- 0,5 kg contre - 0,9 kg/j de lait mesuré sur un mois). Une meilleure production laitière amène un meilleur gain de poids des veaux. Au cours de la lactation entière, le surplus de gain de poids moyen quotidien des veaux est de 60 g/litre de lait bu en plus, soit 70 kg de gain de poids vif pour une lactation de 2 300 kg par rapport à une lactation de 1 200 kg. Une sélection sur la production de la mère et sa persistance permettrait donc de réduire les aliments achetés pour les broutards. Cela conforte la nécessité de pouvoir mesurer la production laitière de façon plus systématique et sur toute la lactation. La production de lait de la vache allaitante est peu sensible à la diminution du niveau des apports énergétiques ou globaux si celle-ci est inférieure à 3 UFL. En effet la fonction de production laitière est priorisée quel que soit l’état d’engraissement au vêlage. Néanmoins, la production laitière peut être réduite significativement lorsqu’une durée de sous-alimentation même modérée se prolonge. Avec des vêlages d’hiver, la qualité de l’herbe ingérée au printemps couvre largement (110%) les besoins alimentaires élevés de lactation. Ceci permet à la vache de relancer sa lactation et de reconstituer les réserves mobilisées en période hivernale et ainsi à l’éleveur d’épargner sur l’alimentation distribuée., Measuring suckler cow Milk Production (MP) consists in the evaluation of the milk drunk by the calf during two repeated weighings before and after suckling. This measurement is difficult to implement and limited to experimental centers. A database was created grouping the milk drunk measured over 15 years by charolais, limousin and salers cattle from the HERBIPÔLE (more than 12,000 measurements, more than 1 000 cows checked at least 7 times per cow and per lactation) in order to update our quantitative knowledge of the production of suckler cows and study the major variation factors. Breed is a major factor of production: 1 600 ± 313 kg milk / lactation in limousin, 1 840 ± 355 kg in Charolais and 2 250 ± 470 kg in Salers and a difference of approximately 11% can be observed between primiparous and multiparous - i.e. 1kg/d. In a “classic” system of “broutard” production with calvings in the winter and semi-extensive pastures, two production peaks are observed. The first peak occurs one month after calving and the second larger peak just after access to pasture. This peak varies from 0.8 to 2.3 kg depending on the breed, parturition and calving period. A focus was made on the factors affecting the peak around access to pasture. The study enabled the quantification of the persistence of lactation and its variability (- 17 to - 27 g milk drunk /d). Persistence is best for low-producing cows (- 0.5 against - 0.9 kg less milk /d over one month). Better MP leads to a better weight gain of the calves. Throughout lactation, the average weight gain is 60 g/liter of extra milk drunk, i.e. a gain in live weight of 70 kg for a lactation of 2 300 kg compared to a lactation of 1 200 kg. Selection of production of the mother and lactation persistence therefore reduces food purchased for weanlings. This supports the need to measure the MP more systematically and throughout lactation. The milk production of a cow suckling her calf is not very sensitive to reductions in energetic or global supplements if the reduction is less than 3 UFL, in fact this function takes priority regardless of the state of weight gain during calving. Nevertheless, the MP can be reduced significantly in the event of a longer period of even moderate under-nourishment. In winter calvings, the quality of grass ingested in the spring largely covers (110%) the high nutritional requirements of lactation. This allows the cow to re-establish lactation and to reconstitute the reserves mobilized during the winter period, and thus the breeder can economize on the feed distributed.

The objective of this work was to determine whether the effects of farming systems on the cheese sensory properties are reduced when the milk used to make cheese is pasteurised. The milk of cows reared under two grazing systems (extensive and semi-extensive) was processed into raw and pasteurised milk cheeses. These grazing systems were mainly differentiable by the milk fatty acid (FA) composition and, to a lesser extent, by their levels of lactic acid bacteria, which were higher in milk from the extensive grazing system. In cheese, the different FA composition persisted and affected the texture of the cheese regardless of the milk treatment. The effects of each grazing system were indistinguishable in several cheese characteristics (proteolysis, texture, and aroma) after pasteurisation. In contrast, among pasteurised milk cheeses, those resulting from an extensive system had greyer rinds, while no difference in rind colour was observed among raw milk cheeses.

Grazing management is an important tool to preserve insect biodiversity. Although literature has discussed the importance of grazing pressure adjustment to support grassland insect communities for the ecosystem services they provide, little has been published on the economic sustainability of such management adjustments to date. This study compared continuous grazing (CG) to an innovative rotational grazing system (the biodiversity-friendly rotation – BR), where a subplot was excluded from grazing for two months during the main flowering period. The effects of grazing two different species (cattle and sheep) within both systems were also evaluated. The aims were to assess the effects on butterfly, bumblebee, and ground beetle assemblages, along with the impact on herbage mass and animal performance. The BR enhanced both the abundance and species richness of flower-visiting insect assemblages and it was observed that cattle provided better results than sheep grazing. A multivariate redundancy analysis highlighted that most of the flower-visiting species (including almost all the endangered and locally rare species) were favoured by BR-cattle treatment, mainly due to the high percentage of flower cover and sward heterogeneity involved in this treatment. However, grazing system and grazer species did not affect ground beetle species richness or abundance. Moreover, herbage mass and animal performance (live weight and body condition score) were comparable between CG and BR throughout the grazing season. The BR could be a useful management system to enhance grassland flower-visiting insect assemblages whilst meeting farm production objectives, especially in protected environments where insect conservation is a major target.

Question(s) Do the effects of grazing and fertilisation on taxonomic and functional diversity follow the dynamic equilibrium model? Location Long-term observatory of permanent upland grasslands in Massif-central, France. Methods Over eight years, we monitored vegetation changes in 24 experimental plots in two productive grasslands after modification of grazing intensity (including abandonment) or fertilisation regime (including cessation of fertilisation). Botanical surveys and plant trait measurements (leaf dry matter content, LDMC; specific leaf area, SLA and reproductive plant height) were performed within each plot. We analysed taxonomic responses using principal response curves and functional responses from observed temporal changes of community-weighted mean traits and Rao index. Results Grazing abandonment led to significant reductions in species richness and functional convergence toward taller plants with higher LDMC in both grasslands. Intermediate and high grazing pressure did not differ but showed strong variations in taxonomic diversity between-years. Changes in functional vegetation structure were mostly driven by species turnover and weakly influenced by intraspecific trait variability. Community responses to modified fertilization regime were more complex with strong differences between the two mown grasslands. Species richness did not change significantly although species evenness decreased in the highly-fertilised treatments (NPK), particularly in the most productive grassland. Here, functional changes were largely driven by intraspecific trait variability. Both functional divergence and convergence were found for different traits. In NPK, grassland assemblages converged toward taller plants while SLA diverged. Conclusions In our upland study system, the different components of grassland diversity responded differently, in terms of direction or rate of change and mostly under extreme modifications of management. Following abandonment, the observed patterns suggest a prominent and consistent role of competitive exclusion as expected under the DEM. However, our findings also suggest the role of additional processes especially the importance of niche differentiation for our mown grassland under NPK. Timing and net effects of these combined assembly drivers were sensitive to even slight differences in initial productivity and species composition. This article is protected by copyright. All rights reserved.

Dans le cadre du projet européen sur l’élevage de précision EU - PLF, l’unité expérimentale Herbip ô le s’est équipée du dispositif CowView (GEA) permettant de suivre en temps réel la position (X, Y) et l’activité de 168 vaches laitières dans son étable de Marcenat (15) . Chaque vach e porte un tag actif (6 GHz) localisé dans le plan 2 fois par seconde et à 30 cm prêt par un réseau de 16 antenne s fixées au plafond (technologie RTLS). Le temps passé par chacune dans différentes activités peut ensuite être calculé, par ex emple l’activité « repos » si la vache est localisé e dans la logette , « alimentation » si elle est localisé e au cornadis .. . Une détection précoce de modification s anormales d e l ’activité d’un individu doit permettre de déclencher des alarmes pour l’éleveur et l’aider dans sa gestion du troupeau. Ce système étant très peu déployé de par le monde (8 fermes équipées) nous avons souhaité le valider puis chercher à l’exploiter pour des besoins de recherches notamment en comportement animal. Durant l’hiver 2015 - 16, nous avons e nregistré les données CowView (1.2Go/j) pour 190 vaches, 24h/24 pendant 123 jours. Pour visualiser et fouiller cet important jeu de données, n ous avons développé un algorithme sous imageJ permettant de convertir l es données de géolocalisation en une image de densité d’occupation , un pixel de l’image représentant le temps cumulé passé par un ou des individus en ce point. Nous avons montré que l’a nalyse de cette image permet tait de valide r les activités des vaches déjà fourni es par le système mais pouvait éga lement révéler de nouvelles activités telles que le temps pass é à se brosser ou encore à lécher la pierre à sel . Une validation a été réalisée par compar aison à un enregistrement vidéo . Elle a montr é que notre démarche était prometteuse et pourrait permett re d’enrichir nos connaissances sur les comportements des animaux en bâtiment.

The aim of this research was to determine the specific influence of milk fat composition on the sensory properties of cheeses derived from cows fed either pasture or a maize silage-based diet. Uncooked pressed Cantal cheeses were manufactured from skimmed milk with an identical chemical and microbial composition and two different pasteurized creams obtained from cows fed either pasture or maize silage. Cheeses produced from the pasture-derived cream were creamier, more elastic, more adhesive, and less firm than cheeses made from the maize silage-derived cream. The rind of pasture-derived cream cheeses was thinner and the relative abundance of Staphylococcus, Brachybacterium et Yaniella were higher than on the rind of the cheeses made with the cream derived from maize silage where Lactococcus was relatively abundant. These results confirm the direct role of fat composition, richer in unsaturated FA with low melting point when cows grazed pasture, on the cheese texture and suggest that the milk fat composition could change the microbiota of cheese rind, which in return can explain differences in cheese rind appearance. The flavour and the microbiota of the core of the cheeses made with the different creams were very similar, indicating that the milk fat composition plays a minor role on the flavour development of cheeses. Other diet-dependent aspects of milk biochemical or microbial composition may explain the stronger and more diverse flavour of cheeses that is frequently reported when cows are fed pasture., Les objectifs de ce travail étaient de comprendre l’influence de la composition de la matière grasse laitière sur les propriétés sensorielles de fromages. Des fromages à pâte pressée non cuite de type Cantal ont été fabriqués avec des laits écrémés qui avaient une composition chimique et microbiologique identique mais qui étaient additionnés de crèmes pasteurisées produites soit par des vaches au pâturage, soit nourries avec de l’ensilage de maïs. Les fromages fabriqués avec la crème des vaches au pâturage avaient une texture plus crémeuse, plus élastique, plus adhésive et moins ferme que les fromages fabriqués avec la crème des vaches nourries avec l’ensilage de maïs. La croûte des fromages fabriqués avec une crème issue du pâturage était plus fine et Staphylococcus, Brachybacterium et Yaniella étaient relativement plus abondant que Lactococcus. Ces résultats confirment l’effet direct de la composition de la matière grasse du lait, plus riche en acides gras insaturés avec un point de fusion bas lorsque les vaches sont au pâturage, sur la texture des fromages. Ils suggèrent que la composition de la matière grasse est susceptible de modifier le microbiote de la croûte du fromage et serait à l’origine de fromages d’aspect différent. La flaveur et le microbiote des fromages fabriqués avec les deux crèmes étaient très voisins, montrant que la composition de la matière grasse joue un rôle mineur dans le développement de la flaveur des fromages. D’autres aspects de la composition du lait expliquerait la flaveur plus diversifiée et plus forte qui est souvent attribuée aux fromages issus du pâturage.

Animal diet quality during the grazing season in two mountain low-input dairy systems. 19. Meeting of the FAO‐CIHEAM Mountain Pastures sub‐network

National audience; "The benefi ts of sedation during thermal dehorning in calves. Dehorning is a painful procedure for calves. The aim of this study was to determine if the combination of asedative and a non-steroidal anti-inflammatory drug(NSAID) could be a means of reducing pain during dehorning and if the of expression pain varied with the age, breed or sex of the calf. Male and female Holstein and Charolais calves aged either 7 or 28 days old were used. Three treatments were studied: dehorning with NSAID(ketoprofen), dehorning with NSAID (ketoprofen) and a sedative (xylazine) or sham dehorning. Calf behaviour, salivary cortisol and cardiac frequency were monitored. The non-sedated calves reacted more strongly and hada higher cardiac frequency and salivary cortisol level than sedated calves in response to dehorning. The responses to dehorning were the same whatever the age, sex or breed of calf. The results indicate that anti-paintreatment, combining sedation and an anti-infl ammatory drug, should be used during dehorning."; "L’écornage est une procédure douloureuse pour les veaux. Le but de cette étude était de déterminer si la combinaison d’un sédatif et d’un anti-inflammatoire non stéroïdien (AINS) est un moyen pour réduire la douleur à l’écornage, et si l’expression de la douleur varie selon l’âge, la race ou le sexe du veau.Des veaux Holstein et Charolais des 2 sexes, âgés de 7 ou 28 jours, ébourgeonnés et ayant reçu un AINS (kétoprofène) seul ou couplé à un sédatif (xylazine) ont été comparés à des veaux sans traitement etpour lesquels l’ébourgeonnage a été simulé. Leur comportement, leur taux de cortisol salivaire et leur fréquence cardiaque ont été suivis.Les veaux non sédatés ont réagi plus fortement et avaient une fréquence cardiaque et un taux de cortisol salivaire plus élevés. Ces effets n’ontpas été observés chez les veaux sédatés. Les réponses ont été les mêmesquels que soient l’âge, le sexe ou la race.Les résultats sont en faveur d’un traitement antidouleur lors de l’écornage, combinant sédation et anti-inflammatoire."

En production de bovins allaitants, l’alimentation du troupeau constitue l’un des facteurs majeurs de la réussite de l’élevage, tant du point de vue technique qu’économique. Elle suppose de bien connaître les besoins des animaux et de bien gérer les ressources fourragères de l’exploitation. Après une présentation des bases de l’alimentation des bovins, cet ouvrage de référence propose une démarche pratique de conduite de l’alimentation du troupeau de vaches allaitantes, mais aussi des veaux avant sevrage et des génisses de renouvellement. Ce livre inclut de nombreux exemples, conseils pratiques, démonstrations pédagogiques et photos. Il a été rédigé par les spécialistes de l’Institut de l’Élevage, avec l’étroite collaboration de l’Inra dans le cadre de l’UMT "Systèmes Allaitants, Fourrages et Environnement" (SAFE). Il s’appuie sur les dernières recommandations de l’Inra en matière d’alimentation des ruminants. Se voulant résolument pratique, cet ouvrage propose des déclinaisons pour les principaux systèmes allaitants français. Il s’adresse à tous ceux concernés par l’alimentation du troupeau bovin allaitant : éleveurs, techniciens, étudiants, enseignants…

Le service de pollinisation rendu par l'écosystème prairial est très souvent mis en avant mais il est peu explicité et peu quantifié. L'objectif de cette étude est de quantifier les interactions plante-insecte à l'origine de ce service dans des prairies appartenant à 3 agrosystèmes contrastés. Pour cela, nous avons capturé les insectes butinant les fleurs sur des transects, à 3 périodes de l'année, dans 18 prairies présentant un gradient de richesse floristique et appartenant à 3 fermes expérimentales INRA située à Marcenat (Cantal), Mirecourt (Vosges) et Lusignan (Vienne). Nous avons testé en parallèle une méthode d'analyse basée sur le méta-barcoding, permettant la détermination simultanée de l'insecte et des pollens qu'il transporte. Nous avons analysé 979 couples fleur-insecte butineur. Les résultats mettent en évidence le rôle important des Diptères dans les prairies permanentes tempérées multi-espèces en tant que butineurs et pollinisateurs potentiels, notamment en début de saison de végétation. Elle montre également que les interactions entre familles d'insectes et familles de plantes observées à Marcenat sont plus nombreuses et réparties plus équitablement qu'à Mirecourt et Lusignan. Une forte proportion des insectes capturés (82 %) transportait du pollen et l'analyse des barcodes ADN révèle que 44 % transportaient en même temps de 2 à 6 genres botaniques différents. Ces premiers résultats permettent d'identifier les acteurs potentiels du service de pollinisation dans les prairies., The pollination service provided by the grasslands is very often highlighted, but it is poorly clarified and poorly quantified. The objective of this study was to quantify plant-insect interactions at the origin of this service in grasslands belonging to three contrasted agrosystems. To do this, we captured insects foraging flowers over 3 periods on transects in 18 grasslands within a gradient of floristic diversity and owned on three experimental farms of Inra located at Marcenat (Cantal), Mirecourt (Vosges) and Lusignan (Vienna). We tested in parallel an analysis method based on barcoding, allowing the simultaneous determination of the insect and pollen it carries. We analyzed 979 flowers-foraging insect couples. The results hightliht the important role of Diptera in the temperate grassland multi-species as foragers and potential pollinators, particularly at the befinning of the growing season. It also shows that interactions between insect families and plant families observed in Marcenat were more numerous and distributed more equitably than in Mirecourt and Lusignan. Moreover, analysis of pollen using DNA barcodes pointed out that a large proportion of the captured insects (82 %) carried pollen and that 44 % carried two to six different botanical genera. These first results allow identifying some actors of the pollination service in grasslands.