Experimental manipulations of ponds and responses.

Carpenter and Kitchell tested their trophic cascade model by manipulating the fish communities in two lakes and using a third lake as a control. Figure 19.17 shows the overall design of their experiment. Two of the lakes contained substantial populations of largemouth bass. A third lake had no bass, due to occasional winterkill, but contained an abundance of planktivorous minnows. The researchers removed 90% of the largemouth bass from one experimental lake and put them into the other. They simultaneously removed 90% of the planktivorous minnows from the second lake and introduced them to the first. They left a reference lake unmanipulated as a control.