The detection of trends in ecosystems depends upon (1) a good description of the foundation or benchmark against which changes are measured and (2) a distinction between natural and anthropogenic changes. Patterns and mechanisms observed over 25 years in a large kelp forest suggest that definition of a meaningful benchmark is impossible, because many of the large animals have been gone for years to decades, and kelps are sensitive to large-scale, low-frequency El Niñ o-Southern Oscillation events and longer term regime shifts. A shift in the oceanographic climate has significantly reduced the average size and carrying capacity of the dominant plant. The animals that have been functionally removed from the community include sea otters, black sea bass, yellowtail, white sea bass, and abalones. Other species are still present, but fisheries have had huge effects on the abundances, size-frequencies, and/or spatial distributions of sheephead, kelp bass, rays, flatfish, rock fish, spiny lobsters, and red sea urchins. Now even sea cucumbers, crabs, and small snails are subject to unregulated fishing. The plants continue to exist without a hint of the effects of the loss of so much animal biomass. Furthermore, most of the megafauna have been removed with very little documentation or historical understanding of what the natural community was like. Thus, our ability to separate anthropogenic impacts from the ''natural'' dynamics of the system is severely compromised. We discuss the importance of both an ecosystem focus on productivity and careful monitoring of as many populations as possible. In addition, we show that this community is not tightly integrated with mutual dependencies; hence, many species can be removed without much affecting the rest of the ecosystem.