10. Plant Diversity in a Patchy Landscape
153
degree of pollen limitation, and (5) survival of transplants on small versus large
outcrops.
Calystegia grows in discrete patches 2 m 2 -300 m 2 in area, consisting of one to
a few genetic individuals. Small outcrops (here, <5 ha) supported 1-2 plant patches,
while large outcrops (here, >300. ha) supported many discrete patches. On large
outcrops, marked ramets produced over twice as many flowers per ramet than on
small outcrops, in both 1995 and 1996 (MANOVA; P < 0.001). Production of seed
capsules was also higher on large outcrops; during 1995 and 1996 respectively,
25% and 27% of marked ramets produced seed capsules compared to 10.8% and
4.3% on small outcrops (MANOV A; P < 0.005). For the 25 plant patches in which
seed capsules were produced (8 on small and 17 on large outcrops), there were no
significant differences in seeds per capsule, seed mass, or proportions of seeds that
germinated (MANOV A; p > 0.5).
Experimental augmentation with pollen from other plant patches led to significantly higher seed production compared with all other treatments; thus, Calystegia
appeared to be self-incompatible and pollen-limited. On a per-flower basis, there
was no difference in the number or diversity of pollinators visiting flowers on small
versus large outcrops. However, production of seeds per flower increased significantly with the number of other plant patches within 100 m. For a given patch,
there were 2.1± 0.5 other patches within 100 m on small outcrops, versus 5.9 ± 1.3
on large outcrops (MANOV A, P < 0.01). Inclusion of this variable removed the
significant effect of large versus small outcrop on seed production. Thus, isolation
from sources of high-quality (= non-self) pollen appeared to be one reason for lower
reproductive success of Calystegia on small outcrops.
5 Microhabitats and the Serpentine Sunflower
With collaborators Amy Wolf and Paul Brodmann, I examined the distribution of
another endemic, the serpentine sunflower Helianthus exilis (Asteraceae). In preliminary surveys in 1994, we found fl. exilis on 13 of 15 large outcrops, in populations of 200--93000 flowering individuals; we found no H. exilis on any of the 30
small outcrops. We undertook experiments and further surveys to explain this striking
pattern of distribution (Wolf and Harrison 1999).
The answer proved to be simple; Helianthus grows in so-called serpentine seeps,
where a spring-fed stream emerging from serpentine crosses a sandy or gravelly
area that traps moisture and remains wet into midsummer. In our regional surveys,
we never found seeps on small outcrops. While this explains H. exilis' absence
from small outcrops, it does not explain any of the results of the first (diversity) part
of this study, because seeps are sufficiently rare that none of my 24 randomly chosen sites on large outcrops included any seep habitat.
Plants that experimentally received extra pollen produced more seeds on average than naturally pollinated controls, suggesting that seed production in Helianthus
is partly limited by pollen. However, we observed no significant relationships be-
153
degree of pollen limitation, and (5) survival of transplants on small versus large
outcrops.
Calystegia grows in discrete patches 2 m 2 -300 m 2 in area, consisting of one to
a few genetic individuals. Small outcrops (here, <5 ha) supported 1-2 plant patches,
while large outcrops (here, >300. ha) supported many discrete patches. On large
outcrops, marked ramets produced over twice as many flowers per ramet than on
small outcrops, in both 1995 and 1996 (MANOVA; P < 0.001). Production of seed
capsules was also higher on large outcrops; during 1995 and 1996 respectively,
25% and 27% of marked ramets produced seed capsules compared to 10.8% and
4.3% on small outcrops (MANOV A; P < 0.005). For the 25 plant patches in which
seed capsules were produced (8 on small and 17 on large outcrops), there were no
significant differences in seeds per capsule, seed mass, or proportions of seeds that
germinated (MANOV A; p > 0.5).
Experimental augmentation with pollen from other plant patches led to significantly higher seed production compared with all other treatments; thus, Calystegia
appeared to be self-incompatible and pollen-limited. On a per-flower basis, there
was no difference in the number or diversity of pollinators visiting flowers on small
versus large outcrops. However, production of seeds per flower increased significantly with the number of other plant patches within 100 m. For a given patch,
there were 2.1± 0.5 other patches within 100 m on small outcrops, versus 5.9 ± 1.3
on large outcrops (MANOV A, P < 0.01). Inclusion of this variable removed the
significant effect of large versus small outcrop on seed production. Thus, isolation
from sources of high-quality (= non-self) pollen appeared to be one reason for lower
reproductive success of Calystegia on small outcrops.
5 Microhabitats and the Serpentine Sunflower
With collaborators Amy Wolf and Paul Brodmann, I examined the distribution of
another endemic, the serpentine sunflower Helianthus exilis (Asteraceae). In preliminary surveys in 1994, we found fl. exilis on 13 of 15 large outcrops, in populations of 200--93000 flowering individuals; we found no H. exilis on any of the 30
small outcrops. We undertook experiments and further surveys to explain this striking
pattern of distribution (Wolf and Harrison 1999).
The answer proved to be simple; Helianthus grows in so-called serpentine seeps,
where a spring-fed stream emerging from serpentine crosses a sandy or gravelly
area that traps moisture and remains wet into midsummer. In our regional surveys,
we never found seeps on small outcrops. While this explains H. exilis' absence
from small outcrops, it does not explain any of the results of the first (diversity) part
of this study, because seeps are sufficiently rare that none of my 24 randomly chosen sites on large outcrops included any seep habitat.
Plants that experimentally received extra pollen produced more seeds on average than naturally pollinated controls, suggesting that seed production in Helianthus
is partly limited by pollen. However, we observed no significant relationships be-