Social connections and cultural aspects play important roles in shaping an individual’s preferences. For instance, people tend to select friends with similar music preferences. Furthermore, preferences and friending are influenced by cultural aspects. Recommender systems may benefit from these phenomena by using knowledge about the nature of social ties to better tailor recommendations to an individual. Focusing on the specifities of music preferences, we study user connections on Last.fm—an online social network for music. We identify those countries whose users are mainly connected within the same country, and those countries that are characterized by cross-country user connections. Strong cross-country connection pairs are typically characterized by similar cultural, historic, or linguistic backgrounds, or geographic proximity. The United States, the United Kingdom, and Russia are identified as countries having a large relative amount of user connections from other countries. Our results contribute to understanding the complexity of social ties and how they are reflected in connection behavior, and are a promising source for advancements of personalized systems.