Al-Zn-Si alloy coatings are widely used for the corrosion protection of steel sheets. In addition to the favorable corrosion properties, the Al-43wt%Zn-1.6wt%Si coatings present several features which are of metallurgical interest. The first one is their surface appearance dominated by the typical spangle morphology exhibiting dull and shiny areas on the coating surface. Further on, the dendrite tips do apparently not grow along the 〈100〉 crystallographic directions that are typical for fcc-metals, but rather along directions in between 〈100〉 and 〈110〉. In addition, continuous variations of the crystallographic orientation, up to 35°, are observed within individual grains of Al-Zn-Si coatings [Sémoroz1 01]. The objective of the present study is to establish a better understanding of the microstructure development which is responsible for the formation of both the characteristic spangle pattern and the important variations of crystallographic orientation within the grains. In order to elucidate these questions, the present study includes three main axes, (i) a detailed microstructure characterization of industrially solidified samples, (ii) modeling work which encompasses microstructure modeling by the phase field method and a geometrical model, as well as the determination of the solid-liquid interfacial energy anisotropy by an inverse method, and (iii) re-solidification experiments aimed at studying the behavior of Al-Zn-Si layers under modified solidification conditions. The results show that the dendrite network spreads quickly in the coating layer at a temperature between 530 to 535°C. During growth, the dendrite tips are separated from the confining boundaries by a thin, solute-rich liquid film. It was found that the preferred dendrite growth directions are in between 〈100〉 and 〈110〉, 28.5° from 〈100〉. Further on, a mathematical expression for the interfacial energy anisotropy of the considered alloy has been determined. The combination of the geometrical model and surface topography measurements allowed concluding that the spangle pattern is due to preferential dendrite growth along one of the two boundaries confining the melt layer. In addition, the new experimental evidence forced to discard the mechanisms previously proposed for the formation of intragranular crystallographic misorientations. The experimental findings acquired during this study indicate that the solidification shrinkage occurring in the area of the grain envelope is the driving force for the formation of the observed intragranular misorientations. The solidification shrinkage leads to the development of tensile stresses in the oxide film covering the coating while it solidifies. These stresses apply on the dendrite network and lead to plastic deformation in the tip area of the growing dendrite arms.