Introduction: Vascular damage comes in many forms with the puncture wound likely the longest known to humans. Experimentally, vascular damage has typically been visualized in mouse models in which there is little bleeding. Under these conditions, damage is limited mostly to the endothelial layer lining the vessel. Visualization has varied from light to scanning electron microscopy. Interpretation has been dominated by intravital microscopy outcomes in which an initial layer of p-selection-exposed, i.e., α-granule secretion-positive, platelets is deposited in association with the damaged vessel wall and extended by the accumulation of a less activated outer layer of platelets. These results have given rise to a Core and Shell model of platelet-rich thrombus formation [Tomaiuolo et al., 2017]. Recently, new mouse models have been presented in which the vessel is punctured to create a 300 to 600 micron wound hole [Tomoiuolo et al., 2019]. Bleeding is now profuse. The puncture wound results have been interpreted within a Core and Shell model. However, two important aspects of the experimental data [Tomoiuolo et al., 2019] suggest that the existing model may not explain the actual results. First, p-selectin expression as a marker for α-granule secretion and platelet activation was present in limited areas towards the periphery of the resulting thrombus, not as well-defined Core. Second, the hemostatic thrombus when viewed at early stages, ex vivo, showed a pebbly distribution of platelet aggregates suggestive of nucleated platelet accumulation rather than the smooth layers that would follow from a Core and Shell model. We hypothesize that nucleated accumulation of platelet aggregates within the puncture hole could provide pedestals upon which localized accumulation of platelets form the infrastructure of a vaulted thrombus whose extravascular capping leads to bleeding cessation. Methods: To test the proposed hypothesis, we visualized the interior and overall structure of the forming puncture wound thrombus in full 3D at sub-platelet level resolution. To achieve this end, we took our proven serial block face scanning electron microscopy (SBF-SEM) protocols for visualizing platelet organelles in 3D [Pokrovskaya et al., 2018] and adapted them to the visualization of forming thrombi over 1000s of image. To localize samples for electron microscopy, we used in vivo antibody labeling [Tomaiuolo et al., 2019]. This approach had the added advantage of enabling correlative light microscopy mapping overall p-selectin, a marker of platelet secretion, and fibrin distributions against 3D, platelet resolution, thrombus morphology. Results: We found that a 1 min thrombus, pre-bleeding cessation, was structured about the localized accumulation of pedestal-like platelet aggregates along the sides of the puncture hole, extended and spaced along the extravascular adventitia. Subsequent pedestal extension formed a "pontoon" bridge that "capped" extravascularly the puncture hole. At 5 min, full bleeding cessation, we found that forming platelet thrombus had a Swiss cheese-like interior of vaults that were continuous with the intravascular vessel lumen and framed by columns of platelets, presumed pedestal extensions. The thrombus was sealed on the extravascular side by a platelet "cap" (Figure). As expected after bleeding cessation, red blood cells accumulated on the intravascular side of the cap. Formation of a tightly sealed cap was dependent on α-granule secretion as indicated by the effect of knockout of VAMP-8, the primary SNARE protein involved in a-granule release. Based upon morphology, vaults within the forming thrombus were lined with apparent procoagulant platelets providing a potential protected surface for coagulation factor activation. Conclusions: We conclude that bleeding cessation in a true puncture wound occurs from the extravascular side of the thrombus rather than through the formation of a platelet plug that fills the hole. We propose an alternative model of bleeding cessation in which localized platelet aggregates are the starting pedestal upon which all subsequent steps in puncture thrombus formation build, i.e., "Cap and Build". The extent to which properties differ among systems remains an open question. Tomaiuolo et al. 2017. Intervent. Cardiol. Clin. 6: 1-12. Pokrovskaya et al. 2018. Blood Adv. 2: 2947-2958 Tomaiuolo et al. 2019. Proc. Natl. Acad. Sci. USA 116:2 243-2252 Figure Disclosures No relevant conflicts of interest to declare.