This article describes a vignette experiment on the effects of temporal embeddedness and network embeddedness in trust situations. The experiment uses a setting in which a buyer wants to buy a used car from a car dealer. We distinguish between effects on trust of the past relation and the effects of the expected future relation between the buyer and the dealer. A buyer can learn about the trustworthiness of the dealer from past transactions of the dealer. Moreover, the buyer can control the dealer if the buyer and the dealer expect more transactions in the future, because the buyer may sanction the dealer if the dealer would act untrustworthy in the present transaction, for example, by refraining from future transactions. Temporal embeddedness facilitates learning and control through the bilateral relation of the buyer and the dealer, while network embeddedness facilitates learning and control through third parties. In the experiment, subjects are asked to compare different settings for buying a used car, while the relation between the buyer and dealer is varied in these settings. We conclude that both learning and control affect trust at the level of the dyad (temporal embeddedness) as well as at the group-level (network embeddedness).