This article presents a new model, the time-based resource-sharing model, which explains working memory spans in adults. The model posits that both the processing and maintenance components of working memory tasks require attention, and that memory traces decay when attention is diverted. The cognitive load of the processing component, determined by the ratio of the number of memory retrievals to the time allowed, influences the maintenance of items to be recalled. The authors conducted seven experiments to test this model, showing that working memory spans vary as a function of cognitive load. The model integrates time constraints and resource-sharing processes, providing a metric for the cognitive load of tasks. It predicts that working memory spans depend primarily on the cognitive cost of the processing component, with the total duration having a moderate effect. The model also suggests that the number of memory retrievals and the time allowed to perform them are crucial factors in determining span performance. The authors compare their time-based resource-sharing model with the task-switching model, highlighting differences in their predictions about the effects of task duration and cognitive load. The results support the resource-sharing hypothesis and provide a new framework for understanding working memory spans.This article presents a new model, the time-based resource-sharing model, which explains working memory spans in adults. The model posits that both the processing and maintenance components of working memory tasks require attention, and that memory traces decay when attention is diverted. The cognitive load of the processing component, determined by the ratio of the number of memory retrievals to the time allowed, influences the maintenance of items to be recalled. The authors conducted seven experiments to test this model, showing that working memory spans vary as a function of cognitive load. The model integrates time constraints and resource-sharing processes, providing a metric for the cognitive load of tasks. It predicts that working memory spans depend primarily on the cognitive cost of the processing component, with the total duration having a moderate effect. The model also suggests that the number of memory retrievals and the time allowed to perform them are crucial factors in determining span performance. The authors compare their time-based resource-sharing model with the task-switching model, highlighting differences in their predictions about the effects of task duration and cognitive load. The results support the resource-sharing hypothesis and provide a new framework for understanding working memory spans.