The hypothesis that changing the fertility level of the substrate would cha
nge the self-thinning line (different slope or intercept) followed by high-
density populations was tested by solving populations of Ocimum basilicum L
. at two densities on a soil-based potting mix adjusted to three fertility
levels (F0, F1 and F2). Fertility level significantly affected the slope of
the thinning lines for both shoot and root biomass. For shoot biomass, mor
e mortality occurred per unit increase in biomass as fertility level declin
ed (the slope of the thinning line became flatter). The slope of the log sh
oot biomass vs. log density relationship was -0.5 at the F2-, zero at the F
1-, and 0.94 at the F0-fertility. For the log root biomass vs. log density
lines, slopes were zero at the F2- and F0-fertility levels, and -0.32 at F1
. Packing of shoot biomass into canopies of individual plants correlated we
ll with observed exponents of self-thinning lines at the F2- and F1-fertili
ty level. Plants at the F2-fertility level required more canopy space to su
pport a given shoot biomass than plants at F1, indicating that shoot compet
ition was more intense at the F2-fertility level for a given biomass. Leaf
area index and size inequality also increased with fertility level for a gi
ven shoot biomass. Density-dependent mortality in populations grown at the
F0-fertility level was highly unusual in having a positive slope for the sh
oot biomass vs. density relationship. Shoot growth per plant was static vs
density declined in the F0-populations; however, root growth per plant incr
eased. All measurements of shoot growth (mass, height, canopy extension, le
af area) remained static in the F0-populations: root mass and length increa
sed in comparison. It is argued that root competition became sufficiently i
ntense to cause the density-dependent mortality seen at the F0-fertility le
vel, with little contribution of shoot competition to mortality. (C) 1999 A
nnals of Botany Company.