Macroscopic fundamental diagrams (MFDs), which relate the total time spent to the total distance traveled, are explored for freeway networks. It is proposed that these macrolevel relations should be observed if the data come from periods when all lanes on all links throughout the network are in either the congested or the uncongested regime. The theory pertains to freeway networks of any size, even when they are inhomogeneously congested and the data are variable in time. Analysis of vehicle trajectories from two freeway stretches of modest physical length supports this theory. Study further reveals that MFDs can be estimated with data from ordinary loop detectors, provided that every link in the network has at least one detector station and that the data are filtered to meet (approximately) the single-regime requirement. Detector data then confirm that well-defined MFDs exist for other freeway stretches and that the relations are reproducible across days. The results demonstrate that the stringent single-regime condition necessary to observe a freeway MFD does arise at times, even if only on shorter-length freeway stretches. The results also explain why previous efforts to observe freeway MFDs without filtering the data have been unsuccessful. Finally, the results suggest that policies to spread congestion evenly over a freeway network can be useful in maximizing the rate that trips are served.