The paper presents a framework and scalable algorithms for tracking short, distinctive phrases that travel relatively intact through online text. The authors focus on identifying and clustering textual variants of such phrases to study the dynamics of the news cycle, which is the daily rhythms in the news media. By analyzing 90 million articles from mainstream media sites and blogs over a three-month period, they observe a typical lag of 2.5 hours between the peaks of attention to a phrase in news media and blogs, with a "heartbeat"-like pattern in the handoff between the two. They also develop a mathematical model to explain the temporal variation in the news cycle, incorporating imitation and recency effects. The study provides a quantitative analysis of the global news cycle and highlights the interplay between news media and blogs in information propagation. Additionally, they find that only 3.5% of the time, stories first appear dominantly in the blogosphere before being picked up by mainstream media. The approach opens new avenues for studying mutation within phrases and the dynamics of information propagation over longer periods.The paper presents a framework and scalable algorithms for tracking short, distinctive phrases that travel relatively intact through online text. The authors focus on identifying and clustering textual variants of such phrases to study the dynamics of the news cycle, which is the daily rhythms in the news media. By analyzing 90 million articles from mainstream media sites and blogs over a three-month period, they observe a typical lag of 2.5 hours between the peaks of attention to a phrase in news media and blogs, with a "heartbeat"-like pattern in the handoff between the two. They also develop a mathematical model to explain the temporal variation in the news cycle, incorporating imitation and recency effects. The study provides a quantitative analysis of the global news cycle and highlights the interplay between news media and blogs in information propagation. Additionally, they find that only 3.5% of the time, stories first appear dominantly in the blogosphere before being picked up by mainstream media. The approach opens new avenues for studying mutation within phrases and the dynamics of information propagation over longer periods.