Species interactions (i.e., neighbor treatment) and habitat had significant effects on total fitness for both M. guttatus and M. laciniatus (Fig. 2, Table 2). However, species interactions primarily affected survival, whereas habitat had significant effects on both survival and fecundity (Figs. 3, 4, Table 2).
Figure 2. Least square means ± se for rank of total fitness, where total fitness is log (fruit mass g + 1) for all experimental individuals.
Download figure to PowerPoint
Figure 4. Least square means ± se for fecundity, calculated as log (fruit mass g + 1), only for those individuals that produced fruit.
Download figure to PowerPoint
Survival in M. guttatus was highly dependent on habitat (F2,304.2 = 8.56, P = 0.0002). M. guttatus individuals in the seep were 3.58 times less likely to survive to reproduction than individuals in the meadow habitat, and 2.98 times less likely than individuals in the stream habitat [Fig. 3; 95% CI for odds ratios: seep/meadow (1.84, 6.99), seep/stream (1.59, 5.59)]. There was no significant difference in survival between the meadow and stream habitats for either species (M. guttatus: t167.9 = 1.44, P = 0.7012; M. laciniatus: t213.2 = −0.29, P = 0.9997). The probability of survival for M. laciniatus did not differ among habitats (F2,183.8=1.05, P = 0.3524). Fecundity followed the same general pattern, with strong habitat effects for M. guttatus, but not M. laciniatus (Fig. 4; M. guttatus: F2,127 = 6.57, P = 0.0019; M. laciniatus: F2,75.5 = 0.99, P = .3782). M. guttatus individuals produced significantly more seed in the meadow than in the stream habitat (t83.3 = 3.48, P = 0.0078).
Overall, interspecific interactions were competitive (total fitness: t507.5 = 3.28, P = 0.0011; survival: t562 = 2.99, P = 0.0029; fecundity: t221.5 = 0.26, P = 0.7930), whereas intraspecific interactions were marginally facilitative (total fitness: t508 =1.88, P = 0.0607; survival: t562 =1.91, P = 0.0572; fecundity: t228.3 =0.23, P = 0.8208) (Fig. 2-4). The effect of species interactions differed among habitats (Table 2). Interspecific competition was more intense in the meadow habitat relative to the stream habitat for fecundity (t216.1 =2.08, P = 0.0388), and marginally so for survival (t562 =1.68, P = 0.0936) and total fitness (t507.9 =1.71, P = 0.0872) (Fig. 2-4). However, the intensity of intraspecific interactions did not differ statistically among habitats for any fitness components (Tables S1–S3). There was no evidence that M. guttatus and M. laciniatus differed in the intensity of either interspecific (total fitness: t509.2 = −0.43, P = 0.6659; survival: t562 = 1.41, P = 0.1587; fecundity: t215.2 = −1.58, P = 0.1149) or intraspecific interactions (total fitness: t509.7 = −0.10, P = 0.9176; survival: t562 = 1.18, P = .2397; fecundity: t215.2 = −1.17, P = 0.2438).
In the seep, there were strong differences in survival between M. guttatus and M. laciniatus (F1, 144 = 19.54, P < 0.0001). M. laciniatus was 10.53 times more likely to survive to reproduction than M. guttatus (95% CI for odds ratio: 3.68, 30.30). Interspecific competition was strong as individuals grown alone were 5.47 times more likely to survive to reproduction than those in interspecific treatments (95% CI for odds ratio: 1.04, 28.81); there was no significant difference between alone and intraspecific treatments. The strength of interspecific competition did not differ between M. guttatus and M. laciniatus (F2,144 = 0.42, P = 0.6574). There were no significant effects of species identity (F1,28.2 = 1.12, P = 0.2993) or neighbor treatment (F2,27.1 = 0.34, P = 0.7149) on fecundity.
In the meadow habitat, the relative performance of either species was dependent on the dominant neighboring species. M. laciniatus with intraspecific neighbors outperformed M. guttatus with interspecific neighbors in both total fitness and survival (total fitness: t157 = 3.79, P = 0.0029; survival: t174 = 4.3, P = 0.0004) (Figs. 2-4). Similarly, M. guttatus with intraspecific neighbors outperformed M. laciniatus with interspecific neighbors in total fitness (t157 = 4.14, P = 0.0008), with a marginally significant difference in survival (t174 = 2.61, P = 0.10). There were no species differences in either total fitness (t158 = .57, P = 0.5712) or survival (t174 = 1.18, P = 0.2398), although M. guttatus did exhibit marginally greater fecundity (t79.6 = 1.84, P = 0.0696). Rather, species interactions had strong effects on total fitness (F2,157 = 15.95, P < 0.0001) and survival (F2,16.92 = 12.96, P = 0.0004). Individuals in intraspecific treatments were 3.86 times more likely to survive to reproduction than individuals grown alone, and 8.47 times more than individuals grown with interspecific competitors [95% CI for odds ratios: intraspecific/alone (1.06, 13.89), intraspecific/interspecific (3.58, 20)]. There was no evidence that the strength of these interactions differed between species for either total fitness (F2,158 = 0.30, P = 0.7412) or survival (F2,174 = 0.94, P = 0.3921).
In the stream habitat, M. laciniatus outperformed M. guttatus in total fitness (F1, 220 = 14.25, P = 0.0002), survival (F1,244 = 14.03, P = 0.0002), and fecundity (F1,102 = 5.46, P = 0.0214). M. laciniatus was 2.9 times more likely to survive to reproduction than M. guttatus (95% CI for odds ratio: 1.66, 5.08), and had greater first year fecundity (t102 = 2.34, P = 0.0214). However, the subset of M. guttatus individuals left to overwinter in this habitat grew larger in the second year than M. laciniatus grows in any of the local habitats (J. Sexton, unpubl. data). Additionally, species interactions affected the probability of survival (F2,244=3.29, P = 0.0388), but not fecundity (F2,104=.73, P = 0.4843). Individuals in intraspecific treatments were 2.2 times more likely to survive to reproduction than individuals in interspecific treatments (95% CI for odds ratio: 1.19, 4.07). The strength of this effect did not differ between species (F2,244 = 0.07, P = 0.9354).
There were significant differences in flowering time among species, with M. laciniatus having a higher probability of flowering in all three habitats (Fig. 5; Seep: χ2 = 25.354, P < 0.0001; Meadow: χ2 = 17.353, P < 0.0001; Stream: χ2 = 30.831, P < 0.0001). Neighbor effects on flowering were significant in all three habitats; the presence of either inter- or intraspecific neighbors increased the probability of flowering relative to treatments without neighbors (Fig. 5; Seep: χ2 = 13.087, P = 0.0014; Meadow: χ2 = 17.998, P = 0.0001; Stream: χ2 = 20.933, P < 0.0001). In the meadow habitat, there was a significant species-by-neighbor treatment interaction, with M. laciniatus flowering earlier in response to neighbors than M. guttatus (χ2 = 10.85, P = 0.0044).
Figure 5. Cumulative proportion of individuals that have flowered through time for each site, species, and neighbor treatment. Arrows indicate date at which individuals were removed from site due to seasonal drying, causing individuals that had not yet flowered to be right-censored. Filled symbols are Mimulus laciniatus, open symbols are Mimulus guttatus. Neighbor treatments are as follows: ■ = alone; ● = intraspecific; ▲= interspecific.
Download figure to PowerPoint