This paper marks the first step in a larger project examining the 1) dynamics of agenda-setting, 2) the dynamics of issue-framing, and 3) the relationship between the two. Here, I present the first batch of findings from a new dataset chronicling all front-page stories in the New York Times over the last two years (January 2004 through February 2006). Each story is coded by topic category and sub-topic category, according to the Policy Agendas Project coding scheme (Baumgartner and Jones 2006). I structure the paper in four parts. First, I give theoretical discussion of agenda-setting dynamics and issue-framing dynamics in the context of relevant literature. Second, I describe the dataset NYT front-page coverage and what these data tell us about which topics have made headline news over the last two years, how these topics vary in the kind and amount of attention they receive, and how attention across and within topics has shifted over time. Common memory of the last two years' of events is sufficient to show that while events clearly matter, they are not the only force driving media attention. Third, I examine agenda-setting dynamics by tracing the number of stories devoted to each topic by week and then analyzing the patterns of change. I find that media attention does not shift gradually between topics from week to week, but moves instead in fits and starts, with changes exhibiting a leptokurtic distribution. This finding supports Baumgartner and Jones' theory of disproportionate information processing - punctuated equilibrium theory - which describes how finite systems such as the public agenda cope with an overload of information not by processing each item steadily in turn but by lurching from one item to another. Fourth, I examine issue-framing dynamics by focusing on the single most dominant sub-topic in the NYT front-page dataset: the Iraq War. Each story about the war is coded by "framing dimension" and, as with my analysis of agenda-setting dynamics, I trace the number of stories devoted to each framing dimension by month and then analyze the patterns of change. Again, I find leptokurtic distributions. This finding suggests that issue-framing, like agenda-setting, follows a pattern of disproportionate information processing. I conclude by offering a theory that will inform subsequent parts of this larger project, that how a topic is framed strongly determines how much attention that topic will receive. ..PAT.-Unpublished Manuscript [ABSTRACT FROM AUTHOR]