Abstract
Experimental populations of S. canadensis were grown from seed at 3 different stem densities for 28 mo. During the 1st growing season increasing density resulted in reductions in leaf-population growth rates, ramet biomass, number of flower heads, and reproductive ''allocation'', and a delay in flowering. There was negligible mortality in response to high density during the first growth season. Clonal growth was also density-dependent; both the length of new rhizomes and the number produced per ramet were reduced with increasing density. As new rhizomes were produced in autumn some emerged to form leafy rosettes and others remained underground. The rosettes had higher survivorship and made a greater contribution to the next season''s ramet population than did the mature underground rhizomes. The overall shoot densities were regulated by a combination of ramet and genet deaths and density-dependent ramet birth within clones. At high densities, genet mortality and clonal expansion of the survivors resulted in a decreasing number but increasing size of genets over time. Populations of S. canadensis grown at 3 densities from intact, interconnected rhizomes showed similar density-dependent growth reductions but the response was smaller. Interconnected ramets may be less sensitive to increased within-species density than new individuals grown for seed. The relationship between ramet weight and density of interconnected ramets did not conform to the ''law of constant final yield'' predicted for populations of individual plants. This also indicates that ecologically, Solidago ramets are not functionally independent ''individuals''.