We conducted a four-year field competition experiment involving Lythru
m salicaria and the North American wetland species that it occasionall
y displaces, Typha angustifolia. The design placed Lythrum and Typha i
n a modified replacement series experiment having four starting densit
ies (64, 36, 16, and 1 plant(s) per quadrat) and Four relative proport
ions of each species (i.e., 0.25, 0.50, 0.75, or 1, [for densities 64,
36 and 16 only]). Results of a repeated measures ANOVA indicate that
overall rate of ramet production (ORP) differed significantly between
species, between density treatments and between years. We also found e
vidence of significant intra-specific competition in both species. In
the first year of the experiment, ORP in Typha was greater than that i
n Lythrum. However, from the second year onward, this situation was re
versed and ORP in Typha was much lower than that in Lythrum. Analysis
of covariance indicates a significant effect of species proportion on
ORP, suggesting that both intra-, and inter-specific competition were
occurring. Log-transformed ratios of input/output (i.e. proportions of
Lythrum/Typha established at tile beginning of the experiment, divide
d by the proportions found at the end of each growing season) demonstr
ated that in the first year Typha gained advantage in the mixtures, wh
ile by the second and third years the output ratio line crossed the 45
degrees threshold, suggesting the two species could coexist. However,
by the fourth year, the output ratio line was entirely above the 45 d
egrees line, suggesting an overall advantage to Lythrum in the mixture
s. The experiment clearly demonstrates temporal development in competi
tive behaviour, and suggests that Lythrum may take at least four years
before becoming competitively dominant in a wetland community. The re
sults also suggest that short-term experiments should be interpreted w
ith care, if they are to be used to forecast the long-term outcome of
competition in nature.