The excavation, in situ and in laboratory, of three Middle Bronze Age hoards discovered in 2011 in the alluvial plain of the Oise at Ribécourt-Dreslincourt (Oise) has provided the opportunity to reflect on these particular Bronze Age sites. The three hoards (F28, F29 and F36) were buried at one end of a small triangular area. Each hoard is located near a post-hole. This disposition has never been observed in France until now. The hoards are unusual as they include objects made from materials such as gold, amber, glass and bronze and some of the objects themselves are unique. They include a hammer with a twisted decoration, a pin with two large discs and a gold disc. The Baltic origin of the amber from hoard F36 is of particular interest (a prestigious material) as well as the nodules of raw material. The variety of objects should be underlined as they fall into four functional categories: weaponry (two daggers), tools (two axes and a hammer), ornaments (eleven bracelets, a pin, three torques, twenty-seven elements strung on a probable necklace) and the gold disc (part of a possible ritual object). Some of these objects appear to have had little or no use, while others are clearly used. Also of note is the presence of miscasts. The disposition of the objects shows rarely observed gestures. In the case of hoard F29, a single metal object is buried, an axe placed vertically, cutting edge upwards. For hoard F28, the stringing of six bracelets on the handle of a dagger stuck vertically into the base of the feature is noteworthy. Concerning hoard F36, the most significant object arrangement is a compact pile of artefacts stacked in the middle of the pit, including a bent pin (36.38) enclosing four bracelets (36.34 to 37) as well as a twisted torque (36.33), itself bent and broken into three fragments. We have also observed pre-depositional manipulations. The state of the objects at the time of their burial is variable: all are complete, but five were manipulated before being deposited. In addition to the cluster of objects described above, the dagger 36.26, which is missing the two rivets that hold its handle, may have been dismantled before burial, while the gold disc, a bipartite object, is missing its bronze support. Hoards F28 and 36A have a particular disposition with an intermediate layer devoid of objects. This could be due to the presence of organic material that has now disappeared or because the pit containing the hoard was reopened at a later date in order to deposit a new set of objects on top. The addition of items could have been planned from the outset as with the multi-phased hoards of Grunty Fen (Cambridgeshire, UK) and Spaxton (Somerset, UK). Objects from the Ribécourt-Dreslincourt hoards, suggest that the three are contemporary, which does not imply that the deposits were simultaneous. The timeline could extend over a century according to the typochronological coherence of the objects. The Ribécourt-Dreslincourt hoards are attributed to the Atlantic Middle Bronze Age 2. The annular ornaments are quite characteristic of this period, as are the palstaves. Although the typology of the other objects is less obvious, it does not contradict this chronological positioning. Radiocarbon analysis also confirms this date (1442-1302 cal BC). The hoards belong to the rather rare category in the Atlantic area of multi-type hoards that bring together different objects, as is the case for the Villers-sur-Authie (Somme) hoard, now disappeared, or the Wylye (Wiltshire, UK) and Hollingbury Hill (Sussex, UK) hoards. The typology of the annular ornaments from the Ribécourt-Dreslincourt hoards shows that the objects come from different areas, the south-east of England, the Armorican Massif, as well as the lower and middle basins of the Seine and Somme rivers. We can also draw a direct parallel between the 36.36 bracelet and anklets from graves in the Lüneburg region (Lower Saxony), or from Schleswig-Holstein in Germany. Comparisons for the pin lead us to southern Germany and Switzerland, while the gold disc is similar to objects from the Nordic Bronze Age area (Denmark, northern Germany and the British Isles). The hoards’ composition sheds light on the movement of individuals, objects and ideas between the south of England and the north of France, but also on relations with other more distant regions (Germany, Switzerland), whereas the amber and glass refer to the Baltic and the Near East regions. To conclude, the fact that posts originally marked the Ribécourt-Dreslincourt hoards shows an underlying intention to make these hoards visible: they were not buried to be concealed. This ostentatious dimension goes hand in hand with the intention of sacralising particular places where interventions subsequent to the deposits would have taken place.