Demography, Ecology and Reproductiv, Biology of California Jewelflower (Caulanthus californicus: Brassicaceae) Final Report to the State of California Department of Fish and Game Contract # FG 146 [ Report Prepared by: Dr. Susan J. Mazer Beth A. Hendrickson, MA Department of Biological Sciences University of California Santa Barbara, CA 93106 Telephone: 805-893-8011 FAX: 805-893-4724 September 15, 1993
150
Embed
Demography, Ecology and Reproductiv, Biology of California ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Demography, Ecology and Reproductiv, Biology of CaliforniaJewelflower (Caulanthus californicus: Brassicaceae)
Final Report to theState of California Department of Fish and Game
Contract # FG 146 [
Report Prepared by:
Dr. Susan J. MazerBeth A. Hendrickson, MA
Department of Biological SciencesUniversity of California
Frequency Distributions of Reproductive Traits ....................................................... 23
Fecundity and factors affecting reproductive output ................................................. 24
Effects of individual plant identity on mean ovule and seed number perfruit ............................................................................................................................ 24
The relationship between plant size and reproduction .............................................. 25
Floral characteristics and reproductive output .......................................................... 26
Giant Kangaroo Rat impacts .i................................................................................... 48Future research ...................................................................................... _................... 49
Appendix C: Data _omfield and g_enhouse studies ............................................... 85
Acknowledgements
We would like to thank several people for their assistance in performing field andlaboratory studies and data analysis for this report. Lorne Wolfe conducted fieldstudies in 1992 and prepared two progress reports on that work. Jason Price assistedin field work in 1992 and 1993, performed much of the greenhouse work, and assistedwith data analysis. Mike Brown, Stephen Brown, Steve Silver, and John Damuth alsoprovided field assistance. Thek Moua conducted the greenhouse pollinations andassistance with seed counts and weights. In addition, we are grateful to Russ Lewisand Roy van de Hoek of the Bureau of Land Management and Dean Taylor ofBiosystems Analysis Inc. for their assistance in locating field sites and for sharingtheir expertise on Caulanthus californicus with us, and to Tony Nelson, ChuckWarner, and George Butterworth of The Nature Conservancy for kindly allowing usthe use of facilities at the Painted Rock Reseach Station on the Carrizo Plain.
September 15, 1993 Mazer/Hendrickson 1
INTRODUCTION
The following is the final report concerning the research we conducted on
California Jewelflower (Caulanthus californieus [S. Watson] Payson, Brassicaceae)
in 1992 and 1993 under contract to the California Department of Fish and Game. In
1986 this plant was classified as an endangered native California species under the
, Native Plant Protection Act by the California Department of Fish and Game. It is
also listed as endangered by the US Fish and Wildlife Service under the Endangered
Species Act.
California Jewelflower historically ranged from the southern end of Fresno
County and the western side of Tulare County to Bakersfield in the San Joaquin
Valley; it was also common in the Carrizo Plain and Cuyama Valley (Taylor and
Davilla, 1986; see Map l). Of 40 known historical locations in this range, Taylor and
Davilla found the species to be extant at only two sites, one of which was an artificial
population (the result of a reintroduction effort). Since 1986, Russ Lewis of the
Bureau of Land Management has located about 30 new populations in the Carrizo
Plain, 19 new populations in the Cuyama Valley and 9 populations on BLM lands in
the Kreyenhagen Hills to the north. Taylor and Davilla report that by 1986
agricultural development had caused the elimination of habitat representing 39% of
the historical populations of this species on the floor of the San Joaquin Valley.
Habitat destruction due to agriculture and grazing appears to be the primary cause for
the decline of the Jewelflower throughout its range. Habitat alteration due to the
invasion of non-native species may also be a contributing factor.
Habitat suitable for California Jewelflower is on gently sloping (generally less
than 20%) sites with loamy sand soils. Russ Lewis has noted a frequent association
of the species with soils that have some quartz content in the sand. Associated plant
communities are juniper woodland in the Cuyama Valley and Kreyenhagen Hills and
Ephedra scrub in the Carrizo Plain, both with understory species mostly consisting of
annual grasses and herbs. Common associated species included goldenbush
red brome (Bromus rubens), and Arabian grass (Schismus arabicus). Caulanthus
californicus tends to occur in strongly aggregated colonies covering relatively smallareas.
September 15, 1993 Mazer/Hendrickson "_
California Jewelflower is an annual plant in the IVIustard family. Seedlings
possess highly characteristic 3-lobed cotyledons, which abscise as the basal rosette
forms. The juvenile stage is represented by a three- to eight-leaved rosette, which
lasts several weeks prior to bolting. When bolting, individuals produce one to several
flowering stems that usually branch from the base of the rosette. Dark red-purple
sepals on buds at the tips of the flowering stems later change to green and white as the
buds open. Flowers are white and pendant, and the fruits are long, flattened
(obcompressed) siliques. (see cover illustration from Abrams, 1944). The spherical
seeds possess no physical structures that aid dispersal; as the fruit dries, it splits
longitudinally to release the seeds below the parent plant.
The principal objective of our study was to provide quantitative data
concerning demographic and reproductive processes occurring within natural
populations. The primary objectives of the field studies in 1992 were:
(1) To quantify variation in l_lantsize within and among populations;
(2) To determine whether seeds produced in the field were viable and easily
germinable;
(3) To detect, if present, easily germinable seeds in the seed bank;
(4) To determine whether pollinator visitation is essential for seed production;
(5) To determine natural levels of potential reproductive success (flower production)
and realized reproductive success (fruit and seed production); and
(6) To quantify variation in and correlations among components of reproduction and
the traits that are correlated with reproductive success (e.g., plant size, flower
size, etc.)
Field work in 1993 included:
(I) A study of seedling demography; and
(2) A study of the effects of possible competition from native and non-native annual
grasses and herbs on survivorship and reproduction.
Greenhouse crosses were performed in 1993 to evaluate the effects of self-pollination,
cross-pollination within a population and cross-pollination between populations on
seed production per fruit and seed mass.
September 15, 1993 Mazer/Hendrickson 3
STUDY SITES
Field investigations in 1992 were focused on five populations in the Carrizo
Plain and Cuyama Valley: Cuyama, Foster Plot, Russ Plot, Population 4, and Travers
Ranch. We used two of the same populations in the 1993 study (Population 4 and
Foster Plot), in addition to another site (Population 6) selected because it contained a
larger population than was available at other sites. This relatively large population
size was necessary for our study of the possible effects of inter-specific competition
on survivorship and reproduction. The major criterion used in 1992 in selecting
populations as study sites was that the adult flowering population must include at
least 80 individuals; in 1993 we selected sites which had at least 200 flowering
individuals in 1992. Two of the poPUlations in the Carrizo Plain (Foster and Russ
Plot) are completely contained within fences designed to exclude large grazing
mammals. These exclosures were established by Bureau of Land Management
biologists (Russ Lewis and Roy van de Hock). The other four populations were not
enclosed and thus the potential existed for greater levels of herbivory by largemammals.
The locations of the sites are described below and are also shown on the map
in Appendix A (Map 2).
1. Cuyama. This site is located on the east side of Santa Barbara Canyon Road
approximately 2.5 miles south of Foothill Road. Just after the second cattle guard on
Santa Barbara Canyon Road, the population can be located by walking about 200
yards to the east towards the river. In 1992, there were 80 flowering individuals in
this site, clustered on a sloping embankment of a dry streambed leading to the river.
2. Foster Plot. This population is located on the south side of Soda Lake Road 15.4
miles from Highway 166 (or 6.0 miles from the south entrance to the Washburn
Ranch). There were 232 flowering individuals in this site in 1992.
3. Russ Plot. This population is located on the west side of Soda Lake Road. 13.4
miles from Highway I66 (or 2.0 miles south of Foster Plot) there is a dirt road
heading south. Approximately 0.3 mile along this road Russ Plot is located about 300
yards to the west. There were 160 flowering individuals in this site in 1992.
September 15, 1993 Mazer/Hendrickson 4
4. Population 4. This population is situated very close to Foster Plot, but it is not
enclosed by a barbed wire fence. There is a narrow dirt road heading southwest from
Soda Lake Road at a junction that is immediately south of Foster Plot. At 0.6 miles to
the west of Soda Lake Road this population is located approximately 400 yards from
the road to the west. There were approximately 200 flowering individuals in this site
in 1992.
5. Travers Ranch. This population is also on the south side of Soda Lake Road close
to the Travers Ranch. At 11.7 miles from Highway t66 take the road heading south
for about 0.5 miles until the end of the fence. The population is on the hillside
approximately 300 yards from the road to the west. There were 108 floweringindividuals in this site in 1992.
6. Population 6. This population is located on the west side of Soda Lake Road. At
16.8 miles north of Highway 166 a small dirt road leads south just before Soda Lake
Road crosses a cattle guard, near a water tank. The road branches almost
immediately, with the right-hand branch following a fence line towards the hills.
After about 0.25 miles on this fight-hand branch, the road and fence make a 90-degree
right turn. The population can be found by walking about 200 yards to the east from
this point towards the crest of the small ridge. Russ Lewis found about 200 floweringindividuals at this site in 1991.
METHODS
1992 Field season
1, _%[JP_LgIL_L.
Transect and plot methods were used to estimate the density of Caulanthus
californicus in both Russ and Foster Plots during the period of peak flowering, March
31-_April 2, 1992. At each site line transects were established at regular intervals and
quadrats were placed at approximately 10 meter intervals along the transect. The
Caulanthus plants within each 1 meter 2 quadrat were counted. Forty quadrats were
sampled at each site.
September 15, 1993 Mazer/Hendrickson 5
2. Demo_raohv of Reproduction.
In order to study- the reproductive biology o_"California Jewelflower, a detailed
examination of marked individual plants was conducted in each of three populations
in 1992 (Cuyama, Foster and Russ). During the period of March 31 - April 2, 100
flowering individuals were randomly chosen in both Russ and Foster Plots. Each
plant was marked with a flagging stake and given an identification number. For each
individual we recorded the following data:
(l) number of leaves; (2) number of reproductive stems bearing flowers; (3) numberof flowers.
There were two types of fruit loss or damage noted in these populations in
1992. Abortion of developing' fruits was recorded if normal development was
arrested or if a fruit was shriveled. Such spontaneous fruit abortion could be due
either to pollinator failure or to the unavailability of resources necessary for fruit
development (even if fertilization of ovules was successful). Second, herbivore
damage was recorded if at least some part of a fruit had obviously been eaten.
Several Lepidoptera larvae were found on the outside of some fruits so this seems to
be a likely cause of at least some of the damage and subsequent fruit abortiori. Fruits
that failed to develop were classed as either "aborted" or "damaged" (a fruit could not
represent both classes simultaneously).
The same individuals were resampled at Foster during the period April 11-i2
and the following measurements recorded: (1) length of the two longest leaves; (2)
plant height; (3) number of developing fruits; (4) number of aborting fruits; and (5)
number of fruits with apparent herbivore damage. During this sampling period these
measurements could not be recorded on plants in Russ Plot because of extremely high
levels of rattlesnake (Crotalis viridis) activity in the study population. On this date 80
individuals were marked at the Cuyama population and the same measurements
recorded as at Foster Plot.
One of the observations made at the onset of flowering of Caulanthus
californicus was that individual flower size was extremely variable. Previous studies
on another species in the same family, wild radish (Raphanus sativus, Brassicaceae),
have demonstrated that traits that set the upper bound to male and female
reproductive success (e.g. pollen and ovule production) are positively correlated with
September 15. 1993 Mazer/Hendrickson 6
flower size (Stanton and Preston, 1988; Mazer and Schick, 1991a, 1991b). Similar
patterns are seen in lpomopsi aggregata (Polemoniaceae) (Campbell, 1992). In
addition, Stanton and Preston (1988) observed in R. sativus that mean corolla
diameter is positively correlatied with the number of insect visits to the flower,
suggesting that large-flowered individuals should both receive and export more pollenthan small-flowered individuals. Other workers have also found that flower size in
wild species is positively correlated with seed number per fruit (Kang and Primack,
1991), and that both decline within individuals as they age. This result implies that, as
resources decline over the course of the reproductive season, both flower size and
reproductive output per flower diminish. It seems likely that the variation in flower
size that we observed in C. californicus both within and among individuals in part
reflects the resource status of the individual plants and flowers.
In order to quantify variation in flower size in C. californicus, one flower from
each plant was collected on April 11 for measurements to be conducted later in the
laboratory. Taking advantage of the fact that flowers mature and open in a very
predictable order within inflorescences of Caulanthus (from the basal positions
• [oldest flowers and developing fruits] towards the apical positions [youngest flowers
and buds] of a flowering stem), the bud representing the next flower scheduled to
open was removed from the primary flowering stem and placed in a small plastic
centrifuge tube and then in a cooler with ice. For each flower bud we measured the
following traits: (1)bud length; (2) ovary length; and (3) anther iength. In
addition, five flower buds were collected from 25 randomly selected plants per site in
order to obtain an estimate of variation in ovule number within and between
individuals.
Final reproductive success was estimated using material collected during a
field trip May 2-4. During this trip at least one fruit was collected from each of the
marked plants in Foster and Russ Plot, and Cuyama. The most recently ripened fruit
that had not yet dehisced was placed in a small envelope. The seeds were counted
and weighed in the laboratory to determine the average weight of individual seeds, an
important determinant of early seedling performance in many plant species. In
addition, three fruits were collected from each of at least 25 plants at Travers,
Population 4 and Cuyama in order to determine the amount of variation in seed
number per fruit within individuals and among plants within a sing!e population.
September 15, 1993 Ivlazer/Hendrickson 7
These seeds are now stored under refrigeration at the University of California (Santa
Barbara), and are available for future cultivation or recolonization efforts.
3. Spontaneous Self-Pollination Experiment.
A study was initiated to determine whether pollinator visitation is required for
fruit set, or whether plants are capable of spontaneous self-pollination without the aid
of insect pollinators.
This experiment was conducted in Population 4 and initiated on April 2, 1992.
Twenty plants that had at least two recently opened inflorescences were randomly
selected. A pair of inflorescences of similar age were selected on each plant. One
inflorescence was left untouched: the control treatment or inflorescence. The other
inflorescence was entirely enclosed within a mesh bag (made from white bridal veil)
and tied off on the bottom with a twist tie: the pollinator excuislon treatment. From
the control inflorescence, the next flower bud scheduled to open was removed and
stored in a plastic micro-centrifuge tube for later measurement. The mesh bag
allowed the inflorescence to exchange gases_but did not allow insects to visit and
pollinate the flowers. If California Jewelflower were completely reliant on
spontaneous self-pollination, then fruit production would be the same on both the
control and bagged inflorescences. If bagge d inflorescences (the pollinator exclusion
treatment) failed to set fruit, then it could be concluded that Caulanthus requires
insect visitation to produce fruits. This would imply that pollinators must be
available during the establishment of new ai'tificial or natural populations of
Caulanthus californicus. Our field work was conducted on windy clays during which
pollinator activity appeared to be absent. In order to determine the taxonomic
identity and typical population densities of those pollinators that successfullly service
C. californicus, future field studies should attempt to observe and to collect
pollinators during windless days or during early morning periods during which wind
velocity is low and pollinators are foraging.
4. Seed Bank Study.
One of the factors that can influence the population size of annual plants is the
presence of a seed bank in the soil. Seed dormancy and the presence of a long-lived
persistent seed bank would be one potential mechanism allowing the long term
September 15, 1993 Mazer/Hendrickson 8
survival of populations growing in the extremely unpredictable desert-like
environment of the Carrizo Plain (Leck et al., 1989). Even if survivorship or
reproduction of germinated individuals is extremely low (or nonexistent) during a
drought year or series of years, dormant seed in the soil may remain viable until
conditions are suitable for germination, increasing the likelihood of long term
population survival and successful reproduction. There is some evidence that
Jewelflower population dynamics are strongly influenced by the behavior and
presence of buried dormant seed. Sites which contained only a few individuals in the
drought years of 1988 and 1991 were found to have hundreds of plants during 1992
surveys after a high rainfall winter (Russ Lewis, CNDDB forms, 1988 to 1992).
Soil samples for the seed bank study were collected on April 11 and 12, 1992
(prior to dispersal of the 1992 seed crop) at Foster Plot, Cuyama, Travers Ranch and
Population 4. At each site 25 flowering individuals were randomly chosen. Two soil
cores (10 cm diameter x 5 cm deep) were removed from within one meter of each
plant. Both cores from each plant were placed in a single plastic bag. The contents of
each bag were air-dried, weighed, and a 250-gram, approximately one-half inch layer
of soil from each bag was then placed on a one-inch layer of sponge rock in a plastic
dish in the University of California Greenhouse. The dishes were watered and
monitored daily for the presence of emerging Caulanthus californicus seedlings in
addition to other herbaceous species. Seedlings emerging from these samples would
represent the seed bank stored during previous years.
A second soil sampling was conducted at Russ and Foster plots on November
15, 1992, prior to the onset of the fall rains. We established two transects traversing
each population, and collected soil cores at two-meter intervals along the transect. A
total of 33 soil samples were collected (16 from Russ plot and 17 from Foster plot)
and placed in separate plastic bags. Each soil sample was air dried and thoroughly
mixed and two 250-gram samples from each bag were spread out on sponge rock in
plastic dishes as above. Half of samples from each transect was put into a cold room
(5 degrees C) for I month for stratification and one set put directly into the
greenhouse. Seedlings emerging from these samples would represent the stored seed
from past dispersal events as well as the seed from the 1992 spring.
September 15, 1993 Mazer/Hendrickson 9
6. Germination studies
Following weighing, seeds collected from field populations were stored in
envelopes in the refrigerator. Approximately 100 seeds that were sown in pots of
University of California arid soil mix and placed under ambient greenhouse
conditions failed to germinate. A variety of treatments was subsequently used to
attempt tO break seed dormancy. These included various methods of scarification to
break the seed coat (the usual site of dormancy mechanisms), stratification (placing
seeds in cold wet conditions for various lengths of time), manipulation of Soil matrix
: potential through varying the relative amounts of sand and soil mix, and different
growth conditions (greenhouse and lathhouse).
Scarification treatments included rubbing seeds with fine (100 grit) sandpaper;
nicking them with a razor blade; and hot water scarification (dropping seeds into
several times their volume of boiling water and then allowing the water to cool).
After scarification, seeds were either planted in pots and watered or placed between
pieces of wet filter paper in closed petri dishes; both pots and petri dishes were then
refrigerated for 11 days. After the cold treatment the petri dish seeds were planted
into pots and all the pots were moved into the greenhouse and watered daily with
rainwater for a period of at least 30 days.
Other seeds were dropped into cold rainwater and allowed to soak for a week
to remove potential germination inhibiting agents. Seeds were then planted into
various soil mixtures, and gibberellic acid (GA3) was added to half the pots in various
concentrations. In order to determine if seeds were viable we sent samples to Ransom
SeedLab forgerminationand seedqualitytests. ,.
1993 Field season
, I. Competition study
One of the potential causes of the decline of Caulanthus californicus
throughout its range in the San Joaquin Valley is competitive exclusion by non-native
species such as Bromus rubens. Taylor (1986) cites a study by Westman et al.
(1985) showing the recent evolution of pollution-tolerant biotypes in Bromus rubens.
He hypothesizes that the puzzling disappearance of Jewelflower during recent
September15.1993 MazedHendrickson10
decades from sites where it had previously persisted in spite of grazing might be clue
to a combination of lack of tolerance for poor air quality and competition with
vigorous, oxidant-resistant Bromus rubens. While we were unable to test the effects
of air pollution on Jewelflower, we did initiate a study to investigate the possible
effects of interspecific competition on the survival and reproduction of the plants.
On February 6, 1993 we established four pairs of 0.25 m2 study plots in the
Foster Plot population. Pairs were matched as closely as possible with respect to the
number of Caulanthus seedlings in each plot and microhabitat characteristics of the
plots. It was intended to establish five pairs of plots, but we could not find a matched
quadrat for the last plot, so it was retained as an additional unmatched control plot. In
each plot we counted the number of Caulanthus seedlings, which on this date were
small rosettes with only a few leaves. Seedlings were marked with toothpicks] The
treatment plot was weeded, removing all plants of other species within the plot and
within about a 10 cm border around the plot. Weeding was done by scraping the dry,
hard soil surface with a knife blade so as to disturb the soil as little as possible. The
plot with the greater number of seedlings was chosen as the interspeeiflc removal
treatment plot, since this would provide a more conservative test of our hypothesis
that the removal of interspecific competition would increase seedling survivorship
and adult reproduction in the Caulanthus plants. If, in spite of higher densities of
Caulanthus, the weeded plots supported larger Caulanthus individuals with higher
survival rates than the unweeded plots, then interspecific competition would clearly
represent an important factor that repressed Caulanthus reproduction and survival.
In a similar manner, between March 4th and 8th we set out 10 pairs of plots at
population 6, and five pairs at population 4.
We returned to the Foster population plots on March 8 and counted survivors
and new recruits. At that time plants were still mostly rosettes, although a few had
started to bolt and produce flower buds. All sites were revisited on March 31 and
April 1, and survivors and recruits were counted. At that time plants were in flower
but most were not yet fruiting. On each plant, we recorded: (1) branch number, (2)
flower number, (3) the number of unopened buds; (4) the number of developing
fruits; and (5) the number of branches on each plant that had been lopped off by
mammalian herbivores. Giant kangaroo rats (Dipodomys ingens) were common at
each site and were probably the main predator, but jackrabbits may also have
September 15, 1993 Mazer/Hendrickson 11
contributed some damage. No cattle or sheep grazing occurred at any of the
populations during our study in 1993.
We returned to 'all the sites on April 18 and May 2, recording survivorship for
each quadrat and the number of branches, branches eaten, flowers, buds, developing
fruits, aborted fruits and aborted flowers for each individual within the quadrats.
2. Greenhouse pollination studies
On January 23, 1993 we visited a site about 0.8 miles NW of population 4,-
about 2/3 of the way between population 4 and population 6. This site was later
called Four-Post Valley, in reference to a group of four fence posts that provide a
landmark visible from Soda Lake Road. Russ Lewis had found 978 plants here in
1991. We removed 20. Caulanthus seedlings from this site, using a bulb planter to
remove a plug of soil containing each seedling and transferring the plug to a three-
inch pot. We brought the seedlings back to UCSB and maintained them in a growth
chamber (14 hr of light/10 hr dark; 20 degrees C day; 15 degrees C night). Additional
seedlings were removed on February 6 at population 4 and at a site just 0.4 miles
south of population 4 which we cNled population 5. Ten seedlings were taken fromeach of these sites.
Plants were kept in the growth chamber until they began to bolt and flower.
At that point they were transferred to the greenhouse and a series of hand-pollinations
were done. Flowers from each plant were hand-pollinated using pollen from either
the same plant or from another plant. Pollen sources were alternated in sequence
along the flowering stem in order to randomize the potential effects of fruit position
(on the stem) on seed set. Following pollination,- each flower was marked with the
identity of the pollen donor using small adhesive tags wrapped loosely around the
pedicel. Crosses were performed using plants from the same population (selfed vs.
outcrossed pollen) and plants from different populations as donors. After the fruitsk
developed they were collected and the number of mature, filled seeds, aborted seeds,
and total seed weight per fruit were recorded. Evidence for inbreeding depression
could include the observation that seeds produced by selfing would be smaller or
exhibit higher abortion rates than seeds produced by outcrossing within or between
populations. Similarly, if outcrossed pollinations between individuals of the same
population were less successful or produced smaller seeds than crosses between
September 15, 1993 Mazer/Hendrickson 12
individuals from different populations, inbreeding depression due to the expression of
deleterious alleles exprssed in homozygous form might be the cause. Outbreeding
depression would be detected by lower seed weights or higher abortion rates
following inter-population outcrosses relative to outcrossed pollinations performed
between individuals derived from the same population.
STATISTICAL ANALYSES
-: We used several statistical techniques to analyze the data collected on
California Jewelflower. First, descriptive statistics were calculated to characterize the
mean and variation (standard deviation, coefficient of variation) of the traits we
measured. Second, the variation around the mean of several important traits was
displayed by plotting frequency distributions. Such illustrations of the data allow one
to view how the data are distributed within a population. Third, we used analysis of
variance to determine whether the three populations studied most intensively (Russ,
Foster and Cuyama) differed with respect to the mean values of several traits. Since
some of the data were not normally distributed, we also used non-parametric tests
which do not assume a normal distribution. These included the Mann-Whitney U-test
for comparing two groups, the Kruskai-Wallis test for comparing three groups, and
the Wilcoxon signed rank test for paired comparisons. Fourth, we used regression
analysis to explore the relationship between plant size (height) and various measures
of reproductive success, as well as between floral characters and reproductive
components.
RESULTS AND DISCUSSION
Observations on population structure and ecolo_,v
Caulanthus californicus exhibits a non-random distribution on two levels:
populations occur in extremely discrete patches or colonies throughout the range of
the species, and the distribution of individuals within a population tends to be
clumped. The tendency for the plants to occur in distinct colonies has been noted by
other investigators (Taylor and Davilla 1986, Jepson 1936). The occurrence of
aggregated colonies of individuals suggests that seed dispersal distance is not very
September 15, 1993 Mazer/Hendrickson 13
great in this species, although it is also possible that the specific environmental
conditions that promote germination or allow seedling survival are aggregated within
natural habitats.
Overall, the density of plants in 1992 was greater in Foster Plot than in Russ
Plot (5.66 vs 2.43 plants/m2). However many of the quadrats in Foster Plot had no
plants in them. Consequently, the variance among quadrats was also much greater in
Foster Plot (32.06 vs 5.93). The high variation in plant density among quadrats
within a population of California Jewelflower is not surprising given the strongly
aggregated distribution of individuals in both populations. The non-random
distribution of individuals (variance-to-mean ratio greater than 1; Ludwig and
Reynolds, 1988) suggests that external microhabitat factors such as an association
with kangaroo rat disturbance or a nurse plant relationship may be controlling plant
distributions within a site. There is a rather striking trend within sites for plant
densities to be highest on and around large mounds. These mounds are most likely
the result of kangaroo rat activity and perhaps mark the site of a dead Ephedra
individual. The mounds also tend to support vigorous populations of Amsinckia,
Lasthenia an d Orthocarpus.
Phenology and reproduction patterns
Phenological observations are mostly based on the 1993 data set, since in
1992 plants were already flowering by the first census date and we had no
information on earlier phenological events such as seedling emergence or the timing
of bolting. Heavy rainfall and.poor road conditions in the winter of 1992-1993
prevented our visiting the sites as early in 1993 as we would have preferred and we
were not able to establish the time of seedling emergence. Many seedlings and
juveniles (small rosettes with 2 to 6 leaves, 1 to 3 cm in diameter) were present at all
the sites we inspected in late January 1993, and seedlings continued to emerge
through February at the Foster population. Seedlings had probably started to emerge
soon after the first heavy fall rains in the Carrizo Plain.
By the first week of March in 1993 most plants were still in the rosette stage,
but some individuals had started to bolt and to form buds. Interestingly, it appeared
that those plants which were bolting tended to be the ones with smaller rosettes: someindividuals with rosettes about l cm in diameter had started to bolt and form a 2 cm
September 15, 1993 Mazer/Hendrickson 14
stalk and tiny buds by March 4 (see Appendix B. Photo 5). Individuals with larger
rosettes tended to remain vegetative longer. Virtually all individuals in the
populations on the Carrizo Plain were in flower and most had some developing fruits
by April 1, 1992, and most were in flower or at least had buds by the same date in
1993, although few plants had developing fruits on that date. Plants appeared to be a
week or two behind in their development in 1993 relative to the same date in 1992.
The largest plants did not have open flowers on April 1, 1993 but had large numbers
of buds. Flowering continued in both years through April, with developing fruits
appearing by mid April during 1993. By mid-April 1993 most of the very small
plants were dead, usually without having produced fruits. By the beginning of May
ripe fruits were present or had already dehisced and flowering was largely over except
in the largest individuals, which continued to produce fresh flowers and buds through
early May in 1993.
Thus, there appeared to be a range of flowering phenologies present within the
population related to plant size, with the smallest plants flowering first and senescing
prior to the time of peak flowering of most individuals, and a few very large plants
starting to flower later in the season and continuing to flower past the peak period for
most of the population. Such a range of size-related flowering times is commonly
observed in annual plants of unpredictable environments (Schmitt, 1983; Dieringer,
1993), although the relative size of early- vs. late-flowering individuals varies among
species. In general, however, the presence of phenologicai variation among
individuals in natural populations probably assures that at least some individuals will
produce fruits even in years where soil moisture is insufficient to allow survival in the
later part of the season. The presence of high levels of variation in the flowering time
among individuals of C. californicus is typical of many annual plant species of desert
We were generally unsuccessful in germinating seed that we collected in 1992
and attempted to grow in the greenhouse and growth chamber. In total, about 500
seeds were subjected to the various treatments, and only one seed germinated during
these tests (it was part of a control group not treated in any way and simply planted in
a sand tray and placed in the lathhouse). This seedling did not survive past the
cotyledon stage. Samples of the seed were sent to Ransom Seed Lab (P.O. Box 300,
Carpinteria, Ca. 93013; telelphone: (805) 684-3427) in 1993. Ransom Lab
September 15, 1993 Mazer/Hendrickson 16
investigators observed little or no germination using different chilling and light
regimes or by removing the mucilaginous coating of the seeds. However, by
"clipping" the seeds (carefully cutting the seed coat with a razor blade to expose the
embryo [while avoiding any damage to the radicle]) -- an event that would occur in
nature only after severe abrasion or recurring temperature and humidity changes --
and adding gibberellic acid (GA3,400 ppm) they were able to force up to 52% of the
seeds to germinate (Table 1). The tetrazolium chloride test for metabolic activity
showed up to 98% staining indicating that almost all the seeds were living, and
germination was 98% for both Foster and Travers Plot seeds using a treatment that
combined clipping, GA3, and chilling for 6 days and light.
Thus, the seed was viable but deeply dormant. It may be that a period of
after-ripening is required as well as appropriate environmental cues for germination.
Another possibility is that there is a requirement for a mycorrhizal symbiont to enable
seedlings to grow in the field (Smith, 1980). Ransom Lab scored seeds as
germinating if the radicle (embryo root) had emerged from the seed coat when the
seeds were placed between moist sheets of unbleached blotter paper in their
incubators. If the seeds we planted in sand and various soil mixes produced a radical
but then died from fungal infections or because they lacked the appropriate
mycorrhizal symbiont we would not have scored them as germinating, since no
cotyledons would have appeared above the soil surface. This could also explain why
the one seedling that did appear did not live very long in the lathhouse.
September 15. 1993 Mazer/Hendrickson 17
Table I. Germination treatments and results from Ransom Seed Lab tests.
SeedTreatment %Germ: %Germ:Foster Travers
5 degrees C for 6 d., then 20 degrees C, light (8 hours) and water.Seeds placed on moist dye-tree blotter paper within 4" x 6" plastic 0 0boxes.
20 degrees C, light (8 hrs.) water. Seeds placed on moist dye-freeblotterpaperwithin4"x6"plasticboxes. I 0
20 degrees C, light (8 hrs.),400 ppmGA3. Seedsplacedon moist dye- 5 0free blotter paper within 4" x 6" plastic boxes.
Tetrazolium chloride. 1%* 92* 98*
Removed mucilage. Seeds placed on moist dye-free blotter paperwithin 4" x 6" plastic boxes. 0 0Dried and rewet. Seeds placed on moist dye-tree blotter paper within4"x6"plasticboxes. 0 0Clipped and 400 ppm GA3. Seeds placed on moist dye-free blotterpaperwithin4"x6"plasticboxes. 52 14
Clipped. 400 ppm GA3, chilled at 5 degrees C tbr 6 d. then 20 degrees, 98 98light (8 hours). Seeds placed on moist dye+free blotter paper within 4"x 6" plastic boxes.* For the tetrazolium chloride test, the percentage figure is the percent of seeds which showed stainin(metabolic activity).
Since the seeds show deep dormancy it is probable that seeds do accumulate in the
field to provide a long-lived seed bank in the soil. This would protect the population
against years in which few or no plants survive to reproduce, since stored seed in the
soil would still be present to produce new plants in subsequent years. This strategy is
common in desert annuals and allows them to persist as dormant populations in an
unpredictable environment (Leck et al., 1989).
1992 Field Study Results
Comt_ari_on of study site_;
The measurements from the marked plants at each of the three populations
(100 plants at Foster and Russ Plots; 80 plants at the Cuyama population) were
examined statistically in order to determine whether populations differed with respect
to various traits. Many reproductive and morphological traits differed significantly
among populations (Table 2). Plant height and mean length of the two longest leaves
were greater in Foster than in Cuyama on April 11. Among flower and fruit traits,
bud length did not differ significantly between the two populations, but both anther
September 15, 1993 Mazer/Hendrickson 18
length and ovary length were significantly greater in Cuyama than in Foster. So,
while overall plant size was greater at Foster than Cuyama on April 11, 1992, the
Cuyama plants had larger floral parts.
In spite of the larger floral parts represented among plants at the Cuyama
sited, the mean number of mature seeds per fruit was significantly lower at Cuyama
than at the other two populations on May 2, 1992 (Table 3). However, seed abortion
rates (detectable within mature fruits) were very low and similar among the three
populations (Cuyama had a lower mean abortion rate but it was not significantly
different from the other two sites). In spite of exhibiting larger stamens and ovaries
than the other two populations, Cuyama plants produced fruits with fewer seeds perfruit.
Seed abortion levels are estimated as the number of shriveled or clearly
inviable seeds found in a mature fruit; however, seeds and ovules aborted very early
in development would probably not be detectable at that point. Aborted seeds visible
within a mature fruit are those that were probably fertilized but failed to complete
development due to a genetic defect or to an inability to garner sufficient resources.
Early ovule and/or seed abortion is probably an important determinant of seed
number, since ovule numbers per ovary are so much higher and less variable than
seed number per fruit. Based on a comparison of the mean number of ovules per
flower collected from randomly sampled plants on April 11 and of the mean number
of seeds per fruit for marked plants on May 2 (Tables 2 and 3) early ovule and/or
seed abortion appears to have been higher at Cuyama than at Foster (45% vs. 31%).
The comparison of fruits sampled from each of about 25 randomly selected
plants at Travers Plot, Population 4, and Cuyama on May 2 again showed that mature
seed number per fruit differed depending on the population. As when the marked
plants were compared, Cuyama plants had significantly fewer seeds per fruit than the
two Carrizo Plain populations (Table 4). Travers Plot and Population 4 did not differ
significantly in seed number per fruit from Russ Plot and Foster Plot (see Table 3), so
all the Carrizo populations were similar with respect to seed number per fruit.
Individual seeds at Cuyama were significantly lighter in weight than seeds
from the other two populations. Foster and Russ populations had seeds with a mean
weight of 0.764 and 0.77 mg respectively and did not differ significantly from each
September 15, 1993 IVlazer/Hendrickson 19
other, while Cuyama population seeds weighed an average of 0.692 mg, significantly
lower than the other two (Fisher PLSD, p = .0099).
In 1992, fecundity (fruit number per plant times seed number per fruit) was
signficantly higher at Foster and Russ plots than at Cuyama; this was a result of
significantly greater number of fruits per plant as well as more seeds per fruit in the
two Carrizo populations. Percent fruit set (fruits produced per flower) was also
significantly higher at Foster plot than at Cuyama. However, the source of fruit loss
differed among sites. The average number of fruits damaged by insect herbivores
measured on May 2 was much greater in Russ (mean = 2.38) than either Foster (0.79)
or Cuyama (0.74; Table 3)i The mean number of aborted fruits per plant was
significantly higher at Foster than at Cuyama on April 11, and higher than either
Cuyama or Russ on May 2 (Tables 2 and 3).
Table 2. Means and standard deviations (in parentheses) for traits sampled on April 11. Russplot was not sampled on that date due to high rattlesnake activity. Ovule number per ovary wasdetermined from a sample of 3 flower buds per plant for 13 randomly selected plants at Fosterplot and 12 plants at Cuyama. To detect statistically significant differences between Foster andCuyama population means, different statistical analysis techniques were applied as warranted bythe data type. These included analysis of variance (Fisher PLSD) and unpaired t-test fornormally distributed data, and the KruskaI-Wallis test and Mann-Whitney U-test for non-normally distributed data. Numbers of plants examined per trait are indicated by the "n"column following each trait. Statistically significant differences between population means areindicated by p-values <0.05; n.s. =non-significantdifference between populationmeans.
Number of aborted fruits 3.9 (4.4) 95 1.5 (I.8) 80 .0001_P'V
Number of aborted flowers 1.51 (3.52) 98 .04 (.25) 80 .0001Mw
V = Fisher_PLSD T = unpaired t-test MW= Mann-Whitney U-test
September t5, t993 Mazer/Hendrickson 20
Table 3. Means and standard deviations (in parentheses) for traits on the May 2 sample date formarked plants in each population. Percent fruit set is the number of fruits present on May 2divided by the total number of flowers, buds, developing fruits, and aborted fruits present onApril 11. Fecundity is the number of fruits per plant times the number of seeds per fruit. P-values less than 0.05 indicate that population means differ significantly. The KruskaI-Wailis testdoes not provide pairwise comparisons between means. For normally distributed traits(indicated by F) means with distinct superscript letters differ significantly at the 0.05 level ofsignificance.
Table 4. Analysis of variance to detect the effect of population on number of seeds per fruit inTravers Plot, Cuyama population, and Population 4 sampled May 2, 1992. Fruits were sampledfrom approximately 25 randomly chosen plants within each population, p-value less than 0.05indicates that there are significant differences among population means with respect to the meannumber of seeds per fruit.
Analysisof VarianceTable
Source: DF: Sum Squares: MeanSquare: F-test:Between,qroups 2 2309.543 [1154.771 3.597
Within,qroups 158 50727_302 ,321.059 p = ,0297Total 160 53036.845
Population No. of fruits Mean# seeds/fruit': Std.Dev.: Std. Error:
Thus there are some clear differences in reproductive traits among populations ..
within a relatively restricted geographical area. Foster and Russ Plots were similar
with respect to fruit and seed production, but the Cuyama population, although only
about 20 miles south of the other two, differed in floral characteristics as well as in
fruit and seed production and seed weight. Foster, Russ, Travers, and Population 4
were all similar to each other with respect to seed number per fruit, and all were
significantly different from the Cuyama population. However, it must be
remembered that these statistical analyses are based on discrete sample dates and the
picture that emerges is like a snapshot of the reproductive output at a particular time:
it is assumed that fruit production did not continue past the last sample date. The
magnitude of the apparent differences in fecundity and floral traits may be partly due
to different timing of the onset of flowering in the different populations.
September 15, 1993 Mazer/Hendrickson 22
Higher elevation populations are usually phenologically less advanced than
low elevation populations, and this may explain the observed differences between the
Cuyama and Carrizo populations. The elevation of the Carrizo populations is about
2200 to 2250 feet; the Cuyama population is approximately at 2700 feet. The
proportion of fresh flowers and buds to the total number of flowers, fruits and buds
per plant was higher in plants sampled at Cuyama on April l I than at Foster (77.8%
vs. 46.4%, p = .0001, Mann-Whitney U-test). This higher proportion of fresh flowers
and buds at the Cuyama population indicates that these plants were still flowering
strongly on April 11, when many of the Foster plants were producing primarily fruits.
Figures 7a and 7b display the frequency distributions for the number of flower buds
per plant on this date at both populations; Cuyama had few plants in the lowest
frequency class (most plants had more than 5 buds). Cuyama plants had larg e
numbers of buds but few developing fruits compared to the plants at Foster Plot on
the same date. Thus the fecundity measures for Cuyama on May 2 may not represent
the true seed production of that population, since the plants might have produced
more fruits that season. However, the lower number of ovules per ovary and seeds
per fruit at Cuyama are probably not attributable to the phenological state of the
population, since ovule production per ovary is generally highest during the early
stages of flower production and declines towards the end of the flowering period
(Mazer, Snow and Stanton, 1986).
Survival
The probability of survival in the period from flowering to fruit maturation
was highly site-specific in 1992 (Table 5). Only 72 of the 100 plants that were
initially marked on March 31 in Foster Plot survived until May 2 to produce seed.
Mortality was highest between April 11 and May 2, with 20 plants dying in this
period. In contrast, survival was higher in Russ Plot (85%) than in Foster Plot
between March 31 and May 2 even though these two sites are only two miles apart. It
is difficult to directly compare the survival during the same period in Cuyama since
plants within this site were marked at a later date. However, there was almost no
mortality in Cuyama population between April ll and May 2. These results
demonstrate that mortality in the endangered Caulanthus californicus varies over
extremely small geographical distances. The low mortality at Cuyama during this
period may have been related to the less advanced phenological state of that
September 15. 1993 Mazer/Hendrickson 23
population (see above), such that plants were not yet stressed to the point of
premature senescence.
Table 5. A comparison of survival and fruit loss by May 2, 1992 among the three mainpopulations of Caulanthus californicus. Aborted Fruits is the percentage of marked plants thathad at least one aborted fruit. Damaged Fruits is the percentage of marked plants that had atleast one damaged fruit. Survival is the percentage of marked plants that survived fromflowering to fruiting. The level of survival for Cuyama is most likely an overestimate sinceplants were marked well after the onset of flowering.
Population Aborted Fruits Damaged Fruits SurvivalFosterPlot 47 11 72
RussPlot 39 28 85
Cuyama 53 15 -98
Frequency Distributions of Reproductive Traits
Plotting raw data in the form of frequency distributions can be an extremely
helpful way to observe patterns within and differences between natural populations.
In Figures 1 through 8 it is clear that the distributions of several traits differ
markedly between populations. Foster Plot and Cuyama populations are compared
since they were dissimilar with respect to their frequency distributions for some traits.
Plant height, which is an estimate of overall plant size, is distributed normally; the
majority of plant individuals are relatively close to the mean, and distributed
symmetrically about the mean (Figures la and lb). In other words, there are similar
numbers of small and large plants while most are intermediate in size. A similar
distribution is seen for seed number per fruit (Figures 4a and 4b). In contrast, flower
and fruit production are highly skewed (Figures 2a, 2b, 3a, and 3b); the majority of
individuals produce relatively few flowers and fruits while _t small number of
individuals account for the majority, of flower, fruit and seed production. This sort of
pattern is extremely common among monocarpic plant species in which the majority
of offspring that constitute the next generation are derived from a nonrandom portion
of the previous year's adult population (Stanton, 1985).
The shape of the frequency distributions can illustrate the difference between
two populations for some of the traits shown in Tables 2 and 3. The distribution of
September 15, 1993 Mazer/Hendrickson 24
plant height at Foster plot appears exhibits a peak at approximately 25 cm, and there
is a long tail of taller individuals to the right of the peak (Figure la). At Cuyama the
distribution is more clustered around the mean of approximately 20 cm, with no very
tail individuals (Figure Ib). Cuyama individuals more frequently had very few fruits
per.plant, and only one plant had a large number of fruits (Figures 2a and 2b). The
range of seed number per fruit was similar in the two populations (Figures 4a and
4b), but more fruits from Cuyama contained lower numbers of seeds. In both
populations there was a peak in ovule number per fruit at about 50 to 60 (Figures 5a
and 5b), but Cuyama has a larger number and proportion of ovaries with fewer
ovules than did Foster Plot.
The frequency distribution of the number of aborted and damaged fruits per
plant was skewed so that most individuals had about one aborted fruit while some
plants had five to ten aborted fruits (Figures 8a and 8b). The percentage of plantswithin the three sites that had at least one aborted or herbivore-damaged fruit in 1992
is presented in Table 5. The probability that an individual suffered some level of
fruit abortion was highest in Cuyama (53%), lowest in Russ (39%) and intermediate
in Foster (47%). However, the number of aborted fruits per plant was highest at
• Foster population, as seen in Tables 1 and 2 above. In contrast, herbivore-damaged
plants were most common in Russ (28%), least in Foster (11%), and intermediate in
Cuyama (15%).
Fecundits, and factors affecting, reproductive output
Effects of individual plant identity on mean ovule and seed number per fruit
We observed a great deal of variation among individuals within California
Jewelflower populations with respect to reproductive traits such as seed number per
fruit, floral traits, and other components of reproduction. Presumably genetic and/or
environmental factors contribute to this variation. Since these factors operate on an
individual plant basis, traits would be expected to be less variable within individuals
than among individuals. We examined the variation within and among individual
plants of two traits relevant to fecundity: ovule number per ovary and seed number
per fruit.
September 15. 1993 Mazer/Hendrickson 25
Ovule number per ovary was determined in up to 5 buds per plant collected on
April t 1 from each of 12 randomly selected individuals at Cuyama and 13 individuals
at Foster Plot. Individual plants at both populations differed significantly from each
other with respect to mean ovule number per ovary (p = .0001, Fisher PLSD, Figure
9). That is, buds from the same plant tended to be more similar to each other in ovule
number than to buds from other plants.
Seed number per fruit was determined in three fruits per plant collected on
May 2, 1992 from 25 randomly selected plants each at Travers, at Cuyama, and at
Population 4. Again, individual plant identity was a significant factor determiniiag
mean seed number per fruit; fruits from the same plant tended to be more similar to
each other with respect to mature seed number than to fruits from different plants
(Table 6, p = .0001, Fisher PLSD).
Table 6. Analysis of variance for the effect of individual plant identity on seed number per fruitfor pooled plants from Travers, Cuyama, and Population 4. Three fruits were collected fromeach randomly selected plant. There were 161 fruits collected from 57 plants total.
One Factor ANOVA X 1 : ID# Y 1 : Seed #
Analysisof VarianceTable
Source: OF: Sum Squares: Mean Square: F-test:Between,qroups 58 33900.511 605.366 3.29
Within ,qroups 104 19136.333 184.003 p = .0001Total 160 53036.845
The relationship between plant size and reproduction
As shown above, ovule and seed number are traits which are relatively
consistent within an individual but differ between individuals. What are some of the
factors that might produce differences among individuals with respect to these and
othei" traits? Habitat variation on a small scale (such as variation in micro-topography
or soil texture or quality) could affect plant vigor, which in turn may affect fruit and
seed production. Genetic differences among plants could also directly or indirectly
result in differences in plant vigor. One measure of plant vigor is plant size.
Regression analysis was used to explore the relationship between plant size and
several reproductive factors.
September 15. 1993 Mazer/Hendrickson 26
Plant height was an extremely reliable predictor of the reproductive output of
individuals of Caulanthus californicus, as shown by the results of the regression
analyses in Figures 10 - 15 (figures show data from Foster Plot except where
important differences exist in the regression for a trait at Cuyama, then both
populations are illustrated). Plants that were relatively tall produced more flowers,
buds, fruits and seeds per fruit (Figures 10 - 13) than shorter plants. The regressions
fit a polynomial curve better than a straight line; this may reflect the fact that plant
volume (including root mass and the ability to garner soil resources) increases
geometrically rather than linearly as plants get larger. Fruit set (the number of fruits
on May 2 divided by the total number of flowers, fruits and buds on April 11) was
independent of plant height (our estimate of plant size). Thus the probability of a
flower developing into a fruit was not dependent on plant height (Figure 14).
Different plants of the same height varied with respect to fruit set (the proportion of
flowers that successfully develop into fruits) from ~0.10 - 0.60 or higher.
There was no relationship between plant size and mean individual seed weight
at Foster Plot (r2 = .045, p = .3131). Mean individual seed weight per fruit was also
independent of the number of mature seeds per fruit in both populations. It seems
that in these plants seed weight (an indicator of seed quality) is conserved even when
stress or other factors cause increased ovule abortion. When resources are limiting
plants produce fruits with fewer seeds but no reduction in seed quality. However
there was a significant positive regression between plant size and mean individual
seed weight at Cuyama (Figure 19, r2 = .176, p = .0009). Thus, at Cuyama larger
plants produced heavier seeds than smaller plants,
Interestingly, the regressions of seed number per fruit on various plant traits
differed between the Foster and Cuyama populations. The Foster plants showed a
much tighter relationships between seed number per fruit and plant height and
between seed number per fruit and leaf length. Among plants from Cuyama, the
correlations, while still significant, had a lower r2 value. This suggests that at Foster
plot resources are closely linked with seed set (Figure 12, r2 = .47), while at Cuyama
the relationship is less predictable (r 2 = 217). At Cuyama, factors other than resource
availability may be more important in determining the number of seeds per fruit at
Cuyama. Other such factors may include pollinator availability or behavior.
Recalling that in 1992 we demonstrated that C. californicus shows little if any
spontaneous self-pollination in the absence of pollinators (see the section on the
pollinator exclusion study), it is reasonable to imagine that differences between
populations in mean seed number per fruit and/or in the relationship between plant
size and seed number per fruit could be mediated by pollinators. If, t'or example, in
the Cuyama population: (a) seed set per fruit is pollen-limited (determined by the
amount of pollen received rather than the amount of maternal plant resources
available), (b) pollinator abundances are lower than at Foster plot, or (c) pollinator
September 15, 1993 Mazer/Hendrickson 30
visitation rates are independent of plant size, then the lower seed number per fruit and
the independence between seed number and plant size exhibited at Cuyama could
reflect differences in pollinator service.
The mean number of seeds per fruit was significantly lower at Cuyama than at
the other two populations (see Table 3). Also, fruit set (the number of whole and
insect-damaged fruits on May 2 divided by the number of flowers, buds, developing
and aborted fruits on April 11) was significantly lower in Cuyama than in Foster plot.
This is evidence supporting the hypothesis that pollinators may limit fruit set to a
higher degree at Cuyama than at Foster. One would expect fruit set to be lower at
Cuyama if pollinators were less abundant than at Foster, since plants apparently do
not spontaneously self-pollinate to any great extent. On the other hand, there was a
higher absolute number of fruit abortions per plant on both April 11 and May 2 at
Foster population (Tables 2 and 3), and the abortion rate (number of aborted fruits
divided by the total number of developing and aborted fruits) was higher at Foster
plot than at Cuyama on both dates (although only significantly so on April 11; see
Table 7 below). This would indicate that some other factor in addition to or instead
of pollinator availability may be responsible for the difference between these two sites
with respect to fruit abortion rates.
The general picture that emerges is reminiscent of the pattern observed with
respect to the effects of herbivory during the 1993 field study (see results in next
section). Plants at Cuyama were shorter than those at Foster plot; at the same time
they had fewer aborted fruits, as did the plants which had suffered herbivory in the
weeded plots. The removal of older fruiting branches by herbivores reduces the
number of developing fruits and aborted fruits per plant. This could also explain the
weaker relationship between plant size and the mean number of seeds per fruit at
Cuyama: shorter plants were not necessarily less vigorous since some of them may
have been short due to clipping of their branches by herbivores. In fact, pruning of
branches has been shown in some species to increase allocation to female functions
(seed production) relative to pollen production (Charnov, 1982). Unfortunately
herbivory due to mammals was not noted in our observations of plants at either site
during the 1992 study, so we cannot evaluate this possible explanation.
I.
September 15, 1993 Mazer/Hendrickson 31
Table 7. Rates of fruit abortion in plants of Foster and Cuyama populations on April It andMay 2 1992. Rate of abortion is defined as the number of aborted fruits on a particular datedivided bv the total number of fruits (developing plus aborted fruits).
Mean rate of fruit abortion on April 11, 1992
Group: No. of plants Meanabort,rate Std.Dev.: Std. Error:
• No significantdifferencesamongmeans(Kruskal-Wallistest, p = .3131)
Spontaneous Self-Pollination
The results of the I992 experimental manipulation to determine if C.
californicus is capable of setting seed in the absence of pollinators yielded highly
significant results (Table 8). On average, inflorescences that were enclosed in mesh
bags produced less than one fruit. In contrast, control inflorescences produced
approximately 7 fruits each. Thus, reproductive success in this species is highly
dependent on the activity of pollinators.
Despite bearing relatively showy purple and white flowers, Caulanthus
populations did not attract many pollinators during the day. In fact, only two
bumblebee (Bombus) individuals were observed visiting flowers during the entire
1992 study: one was seen in Population 4 and the other was observed at Foster.
However, given that the plants did not readily self-pollinate when insect pollinators
were excluded, the insects were probably there but were not seen due to the timing of
September 15, 1993 Mazer/Hendrickson 32
observations. That is, pollinators active primarily during the early morning and
evening hours would likely not be observed by investigators present during the
middle part of the day.
Table 8. Results of the pollinatorexclusionexperiment. Valuesare the mean:1:s.d. of thenumberof fruits producedby two differentgroupsof inflorescences.See MethodsSectionfordetails. The twomeansaresignificantlydifferentfromeachother(t=7.12,P<<.001).
Treatment Mean # developing fruits+ s.d. Sample size
14,3- 1.3,1 1.073 1.097 16797,I1,Minimum: Maximum: Ran,qe: Sum: Sum of Sqr.: # Missinq:
I0 I1 I1 18.342I,.,08 12* Significantly different means (Mann-Whitney U-test, p = ,0364)
Causes of mortality in weeded and control plots
We could not obtain reliable estimates of the number of plants removed by
herbivores vs. the number dying of other causes, however the degree of herbivory on
the surviving plants suggests herbivory may have been a major cause of mortality for
those plants which disappeared between sampling dates. On April 18, 52% (81 of
156) of the plants in the weeded plots had suffered herbivory with a mean of 2.8
branches eaten per plant, while only 29% (50 of 174) of the plants in the control plots
had suffered herbivory, with a mean of 1.7 branches eaten per plant (Tables 12 and
13). These differences between treatments were significant (Mann-Whitney U-test,
p=.005). Thus plants in weeded plots experienced herbivory more often and more
severely than plants in control plots. This was confirmed by observations of plants in
the weeded plots sometimes being virtually chewed down to the ground level, while
plants in adjacent control plots were almost untouched (for photos, see Appendix B).
September 15, 1993 Mazer/Hendrickson 35
Table 12. Comparison of the mean number of branches eaten per herbivore-damaged plant inweeded and control plots (only plants with one or more branches eaten are included_. All threestudy sites are combined.
It should be noted that out of 89 non-reproductive plants which died between
the sample dates at the beginning and the end of March, the remains of only 2 were
found and could be determined to have died from some sort of environmental stress.
The other 87 simply disappeared, suggesting loss through laerbivory. Between March
31 and April I8, there was apparently more stress-related mortality, with the remains
of 35 dead (dessicated) plants still present out of 73 deaths. Out of the 106 plants
which died between April 18 and May 2, the remains of 31 were still present. The
apparently greater losses to herbivory for plants in the rosette stages (prior to March
31) may simply be due to the fact that it is more difficult to locate dead rosettes than
September 15, 1993 Mazer/Hcndrickson 36
plants which were reproducing prior to death (rosettes are more likely to "disappear"
than are adult plants, regardless of the cause of death).
There was no significant effect of weeding on stress-related mortality detected
on .April 18; standing dead plants were equally common in weeded and in control
plots. Plants which died and remained in the plots tended to be the smallest plants in
the plots. Usually these were plants with one stem less than 10 cm tall and with no
more than one or two flowers and a few small buds. None of these plants appeared to
have produced any fruits. Thus differences in mortality among plots can be assumed
to be due to greater herbivory rates and/or higher stress in the weeded plots. Higher
levels of interspecifie competition in the control plots did not result in the
expected reduction in survivorship and reproduction.
Comparisons of the three populations showed there was a population effect on
the mortality rate in weeded plots in March. On March 31, the mortality rate in
weeded plots at Population 4 was 59.3%, while in the other two populations (Foster
and population six) the rates were 8.3 and 13.5% (Table 14). Herbivore activity was
extremely high at population 4 during this period (for photos, see Appendix B).
Cumulative mortality in weeded plots did not differ significantly among populations.
Mortality in control plots did not differ significantly among populations for any
census interval or cumulatively.
September 15, 1993 Mazer/Hendrickson 37
Table 14. Comparison of mortality rates in weeded plots in three populations (Foster,Population 4, and Population 6) during March 1993. The Kruskal-Wallis test indicates thatthere were significant differences among populations with respect to the mean mortality rateobserved within the weeded plots.
Kruskal-Wallis X 1 : POP Y 2 : mort r_atoMar
13: 2
# Groups 3#Cases 19
H 8.324 _p= .0156
H correctedforties 8.596 p = .0136
# tiedgroups 2
Population No.ofplotpaim Meanmob.rate Std.Dev.: Std.Erro_
Foster 4 .083 .167 .083
Four 5 .593 .339 .151
Six 10 .135 , .138 .043
Reproductive components in weeded and control plots
As described above, herbivory in weeded plots was much greater than in
control plots. This produced effects on reproductive components, since the
herbivores completely removed flowering branches. Over all populations, the
number of branches removed was higher in weeded than in control plots when we
compared plot means on theApril 18 sample date (see Table 15). The mean number
of whole (intact) branches per plant per plot was significantly lower in weeded plots,
while the mean number of branches eaten by herbivores was significantly higher. The
mean number of developing fruits per plant was significantly lower in the weeded
than in the control plots, despite the fact that fruit and flower abortion was
significantly higher in the control plots. This last effect can be explained by the
higher herbivory rate in weeded plots, where mature flowering br_inches with greater
numbers of fruits, aborted fruits and flowers appeared to be preferentially removed by
herbivores. Plants which hadbranches eaten early in the season produced new
flowering stems rapidly, so even though plants in the weeded plots had more branches
September 15, 1993 Mazer/Hendrickson 38
eaten and fewer branches, the total number of flowers and buds cumulatively
produced per plant was roughly equal to those in the control plots. The trends within
each population were similar to that in the pooled set of data; however, within Foster
plot and Population 4 differences were not statistically significant between treatments,
although the number of plots was small.
Table 15. Comparison of trait means between weeded plots (n = 18plots) and control plots (n =18 plots) over all populations on April 18, 1993. Numbers in parentheses are standarddeviations; p-values are reported for significantly different means (Wilcoxonsigned rank test).
Traits Weeded Mean Control Mean p
Number of whole branchesperplant 1.09 (.77) 1.67 (I. 13) .0097
Number of branches eaten per plant 1.45 (1.54) .617 (.757) .021
Number of flowers per plant 5.20 (6.54) 5.10 (6.49) n.s.
Number of buds per plant 10.58 (10.44) 8.49 (7.29) n.s.
Number of developin[ fruits per plant 2.92 (5.58) 5.32 (7.64) .0052
Number of abortedfruitsperplant .30 (.73) 1.25 (1.65) .0047
Number of abortedflowers per plant .28 (.50) 1.30 (1.75) .0219
Qualitatively similar patterns were seen on the May 2 1993 sample date. The
number of branches per plant was roughly equal between weeded and control plots,
probably due to the ability of the damaged plants to produce new stems very rapidly
(Table 16). The number of flowers and buds per plant tended to be higher on the
weeded plants, while the number of developing fruits tended to be lower. Aborted
fruits and flowers were again fewer on the plants in weeded than control plots,
although this was significant only for aborted flowers.
September 15, 1993 Mazer/Hendrickson 39
Table 16. Comparison of trait means between weeded plots (n = 16 plots) and control plots (n =16 plots) over all populations on May 2, 1993. Numbers in parentheses are standard deviations;p-values are reported for significantly different means (Wilcoxon signed rank test).
Traits Weeded Mean Control Mean p
Number of whole branches per plant 2.17 (1.67) 2.15 (1.75) n.s.
Number of branches eaten per plant 3_91 (3.00) 1.51 (1.28) .0061
* Significantly different means (Mann-Whitney U-test, p = .0388
Table 18. Comparison of mean seed abortion rates for within-site vs. between-site crosses.Within-site crosses include only crosses between individuals (self-pollinations are not included inthis comparison).
Group: No. of fruits Meanseedabort, rate Sld. Dev.: Std.Error:
within site 154 .085 ,2 .016betweensites '34 .1761 .291 ,05
It was observed that seed number per fruit in all the greenhouse-grown
individuals was much lower than that in field-collected fruits regardless of the
pollination treatment (compare Tables 3 and 17). Mean individual seed weights
were also somewhat lower, especially when compared to the Carrizo populations
where these greenhouse plants originated. Plants grown in the greenhouse (where
light levels are low compared to the field) appeared etiolated, and flowers were small
compared to field-grown plants. This observation suggests that. environmental
variation in resource levels in the field may account for the observed variation in
flower size in the field.
Several factors may account for the apparently lower resource status of
greenhouse-raised plants. First, lack of sunlight in the greenhouse could be partially
responsible for the relatively low number of seeds per fruit and a low seed weight.
That is, low seed number per fruit could be due to the production in the greenhhouse
September 15. 1993 Mazer/Hendrickson 44
of relatively small flowers with small ovaries and few ovules. In addition, rates of
early ovule and seed abortion may have been higher in the greenhouse than in the
field, resulting in reduced seed production per fruit in the greenhouse. Second, the
relatively small pot size of the cultivated plants (plants were raised in the greenhouse
in a-inch pots) may have contributed to their small size, Third, the seedlings
collected from the field and raised in the growth chamber experienced long days
relative to the concurrent daylengths in the field. The growth chamber may have
promoted precocious flowering before the rate of root growth and/or of
photoassimilation had reached levels that normally promote or allow bolting in the
field. Such precocious flowering may have contributed to the production of the small
flowers, low seed numbers, and small seeds observed in the greenhouse.
SUMMARY AND MANAGEMENT RECOMMENDATIONS
Populations of California Jewelflower (Caulanthus californicus) in the
Carrizo Plain and Cuyama Valley were found to occupy relatively small, discrete
areas. Individuals within the populations exhibit a clustered or aggregated
distribution. Since there appears to be little seed dispersal, it is possible that
populations of California Jewelflower have always been relatively small and largely if
not totally reproductively isolated, leading to the formation of populations highly
adapted to a particular site and to frequent inbreeding. There appeared to be some
degree of reproductive isolation between relatively close populations (outbreeding
depression), and also some inbreeding depression when plants were selfed. Crosses
that produced the highest quality (largest) seeds were outcrossed pollinations between
individuals from the same population. There was little spontaneous self-pollination in
the field in the absence of pollinators, so the breeding system can be assumed to
involve outcrossing to some degree. More information is needed about the long-term
effects of inbreeding and outbreeding depression in this species, particularly if
revegetation efforts include the establishment of new populations from seed sources
of limited genetic variation. Owing to time and personnel constraints, we wereunable to determine whether the differences in seed. mass we observed between selfed
and outcrossed pollinations resulted in differences in progeny performance following
germination.
September 15, 1993 Mazer/Hendrickson 45
Seeds of California Jewelflower showed a high degree of dormancy, and the
formation and persistence of a seed bank is probably a key factor enabling these small
populations to maintain their genetic diversity and to persist in an unpredictable
environment (Leck et al., 1989). The living population is likely to be both much
larger and more genetically diverse than is apparent from the aboveground population
because the soil seed bank may persist for years and contain many generations ofseeds.
California Jewel flower seedlings appeared to emerge following the first heavy
winter rains in December, and remained in a rosette stage until March, when
flowering began. Initiation of flowering varied among individuals, with some very
small plants bolting at the beginning of March and some of the largest plants still not
having open flowers by April i. Variation in flowering phenology could be related to
genetic or environmental factors, and probably serves to ensure that at least some
(early flowering) individuals are able to set fruit in years when the soils dry out early,
while in more favorable years large plants that delay flowering are able to produce
particularly large numbers of seeds.
Populations of California Jewelflower showed within- and between-site.)
variation in a number of floral and reproductwe traits. Populations in the Carrizo
Plain were similar to each other with respect '_omost traits and all differed from the
Cuyama Valley population in a number of floral traits and in reproductive output.
Some of the differences may have been due to the slightly later flowering Of the
higher elevation population at Cuyama. Other potential causes for differences
observed between populations with respect to reproductive output include geographic
variation in the composition, abundance, or behavior of insect and mammalian
herbivores and/or the abundance and behavior of pollinators. Abiotic environmental
differences between sites may also contribute to the observed differences between
populations in these two areas.
Mana_,ement considerations
If reintroduction of the species to sites within its historical range is to be
considered as one of the components of a recovery plan, the above observations raise
important considerations. Reintroductions may fail if:
September 15, 1993 Mazer/Hendrickson 46
1. The site selected does not possess the particular abiotic (soil, aspect, temperatures
etc.) and biotic (pollinators) habitat characteristics required for species to establish
or to reproduce;
2..The introduced population originates from a site which is far away from the
reintroduction site, and which is characterized by a distinct temperature and/or
rainfall regime. The environmental cues necessary for germination or flowering
of the introduced population may not exist at the new site.
3. The genetic diversity of the introduced seedbank is insufficient, so that inbreeding
reduces plant performance over time;
4. The introduced seedbank is not sufficiently diverse to produce a population
capable of adapting to the new site and surviving there over the long term; and
5. The introduced seedbank contains individuals from widely separated populations
and the resulting offspring suffer from the effects of outbreeding depression.
One strategy to minimize the potential traps of inbreeding and outbreeding
depression and to increase the chances of producing a population diverse enough to
allow adaptation to a site would be to collect seed from several populations as near as
possible to the reintroduction site. The populations should all be relatively close
together (within a mile or less if possible). If clear morphological differences
between populations are observed, this may indicate strong genetic divergence of
populations. In this case, to avoid possible outbreeding depression, test-crosses
between individuals from the alternative populations may be conducted to determine
whether they are genetically compatible. Alternatively, seed from only one
population source could be used for each colonization trial.
Compromises will be inevitable when planning a seed collection strategy prior
to recolonization or revegetation efforts with C. californicus. If more than one
population is used as a seed source, one risks some level of genetic incompatibility
(outbreeding depression) following matings in the field of individuals representing
distinct populations. However, since seeds from several sites would probably be
genetically more diverse than seed from a single source, the probability that some
seed would be adapted to the uncolonized site may be increased relative to single-
September 15. 1993 Mazer/Hendrickson 47
source seed collections. If the strategy used aims to collect seeds that are as
genetically diverse as possible, then habitats from which seeds are collected should be
as different as possible with respect to slope, aspect, soil type and vegetation: Seed
from each population should be collected over a period of 3 - 5 years and stored until
use, in order to sample the diversity present in the seedbank of each source
population. Only one fruit per individual should be taken per year, and as many
different individuals as possible should be sampled within each population (while
avoiding inappropriate impacts to the population by collecting no more than 5% of
the seeds from any particular site in any one year). Test crosses between individuals
representing distinct habitats can be conducted in the field (transporting pollen
between sites in glass vials) or in the greenhouse to determine the level (if present) of
genetic incompatibility between populations.
lnterspecific interactions
Although our study provided no direct evidence for interspecific competition
having a negative effect on survival or reproduction, our observations suggest that it
may play a factor in some situations. Introduced annual grasses such as Brornus
rubens were present at all the study sites, but did not dominate any of them as Bromus
rubens does at some of the historical sites for this species in the San Joaquin Valley
visited by Taylor and Davilla (1989). In one of the control plots where there was a
particularly high cover of annual grasses many of the Caulanthus plants died after
bolting but before producing any fruit. Most of the other C. californicus plants
appeared to be stressed, and it is possible that this was a result of competition with the
grasses.
It should also be noted that the effects of interspecific competition may not be
highly detrimental in high rainfall years such as the two years of this study.
Resources would tend to be more limiting in a low rainfall period, and that is when
competitive interactions are more likely to be important.
Removal of potential interspecific competitors by weeding did produce a
substantial and unexpected effect due to the predation of branches and fruits by
mammalian herbivores (giant kangaroo rats and rabbits) on the weeded plots. Plants
which were in the unweeded plots had lower mortality and lower probability of
predation, probably because they were less obvious to the animals. They also had
September 15. 1993 Mazer/Hendrickson 48
greater reproduction, since herbivores removed entire flowering branches as they
reached maturity in both the control and weeded plots, but they did this with greater
frequency in the weeded plots.
Although we were not able to directly investigate the presence or effects of
interspecific competition in California Jewelflower populations our results do have
some implications for other aspects of management. Grazing by cattle or sheep
during the rosette stage of the plants is unlikely to result in much damage by direct
herbivory; however if the overall plant cover is reduced by grazing then bolting
Jewelflowers will be more obvious and subject to herbivory from small mammals and
cattle. Herbivores can effectively prevent fruit production by repeated grazing, since
usually the entire inforescence is removed at once and a new flowering branch must
be initiated from the base of the plant, which in turn is then removed as it matures.
Heavy grazing, particularly between February and mid-May, is likely to be
detrimental to California Jewel flower populations. Light grazing during the period
before the plants have bolted (bolting may start between late February and April,
depending on the year and the site) may be an acceptable practice, but more
information is needed on the impacts of large grazers before guidelines can be set. In
particular, the effects of grazers on soil texture (due to the compacting effects of
heavy animals) and subsequent germination and seedling survivorship must beassessed.
Giant Kangaroo Rat impacts
Co-management for this and other endangered species in the Carrizo Plain
raises some interesting problems. Our weeding study indicated that small mammal
herbivores can have a substantial impact on California Jewelflower reproduction by
repeatedly removing inflorescences as they mature, retarding or even preventing the .
formation of mature fruits by some plants. One of these mammalian predators is the
giant kangaroo rat, itself an endangered species and probably one of the principal
herbivores on California Jewelflower at the Carrizo Plain sites. Kangaroo rat
populations appeared to be increasing rapidly over the last two years and impacts to
Jewelflower populations may become serious if the area is overpopulated by the
kangaroo rats on a long-term basis. Increasing densities of predators on kangaroo rats
and the normal fluctuation of small mammal population sizes will probably alleviate
September 15, 1993 Mazer/Hendrickson 49
such problems in the long term if the natural ecological balance is maintained in the
system.
Future research
Several questions raised by this study deserve closer attention in future research with
this species. Projects could include:
1. Conduct detailed studies of the abiotic environmental factors that permit
successful persistence of Caulanthus califon_icus populations. What do all sites
share with respect to: yearly mean temperature; temperature maxima and minima;
rainfall; soil type; soil texture: slope and aspect?
2. Identify primary pollinators of Caulanthus californicus to ensure that new sites of
population establishment will contain sufficiently abundant natural or introduced
populations of them.
3. Cultivate Caulanthus californicus colonies in the greenhouse to produce an
experimental seed bank. Pilot studies to determine the ability to establish new
colonies using these cultivated seeds should be undertaken. The success of this
effort would reduce the need to remove seeds produced in the field to provide a
seed source for new populations.
4. Conduct more thorough and systematic sampling of flowers to determine if there
is actually dimorphism in flower type; and, if so, whether the two types of flowers
occur within an individual or on different plants.
5. Compare outcrossing rates and male and female reproductive success between the
two types of flowers.
6. Examine potential environmental and genetic correlates of floral dimorphism.
7. Perform further outcrossing experiments using plants from more distantly related
populations (eg. Cuyama, Kreyenhagen Hills, and Carrizo Plain populations) to
determine whether distant outcr0sses consistently result in increased or decreased
September 15, 1993 Mazer/Hendrickson 50
performance. Other measuresof offspring fitness should be employed as well as
seed counts and weights: germination rates, establishment, growth rates and
reproduction of offspring should all be measured.
8..Conduct greenhouse studies to determine potential interference of Bromus rt,bens
on California Jewelflower, and possible interactions with air pollutants.
September 15. 1993 Mazer/Hendrickson 51
BIBLIOGRAPHY
Abrams, L. 1944. Illustrated flora of the Pacific States. Stanford University Press,Stanford, CA.
Barrett, S. C. H. and J. R. Kohn. 1991. Genetic and evolutionary consequences ofsmall population size in plants: Implications for conservation. In: Geneticsand Conservation of Rare Plants. D. A. Falk and K. E. Holsinger (Eds.).Oxford University Press, Oxford.
Campbell, D. R. 1992. Variation in sex allocation and floral morphology inIpomopsis aggregata (Potemoniaceae). American Journal of Botany. 79:516 - 521.
Charnov, E. L. 1982. The Theory of Sex Allocation. Monographs in PopulationBiology. Princeton University Press, Princeton, New Jersey.
Dieringer, G. 1991. Variation in individual flowering time and reproductive successofAgalinisstrictifolia (Scrophulariaceae). American Journal of Botany. 78:497 - 503.
Elias, T. S. (Ed.) 1987. Consen'ation and Management of Rare and EndangeredPlants. California Native Plant Society, Sacramento, CA.
Fox, G. A. 1989. Consequences of flowering time variation in a desert annual:adaptation and history. Ecology 70:i294 - 1306.
Fox, G. A. 1990a. Components of flowering time vaariation in a desert annual.Evolution 44: 1404- 1423.
Fox, G. A. 1990b. Drought and the evolution of flowering time in desert annuals.American Journal of Botany 77:1508 - 1518.
Hoover, R. F. 1970. The Vascular Plants of San Luis Obispo County, California.University of California Press, Berkeley CA.
Kang, H. and R. B. Primack. 199 I. Temporal variation of flower and fruit size inrelation to seed yield in Celandine poppy (Chelidonium majus, Papaveraceae).American Journal of Botany. 78:711 - 722.
Lacey, E. 1986. Onset of reproduction in plants: size- versus age-dependency.Trends in Ecology and Evolution 1:72 - 75.
Leck, M. A., V. T. Parker, and R. L. Simpson. 1989. Ecology of Soil Seed Banks.Academic Press, San Diego, CA.
Lewis, R. I988, 1991. California Natural Diversity Data Base forms describing newCaulanthus californicus populations in the Carrizo Plain, Cuyama Valley andKreyenhagen Hills.
Ludwig, J. A. and J. F. Reynolds. I988. Statistical Ecology. John Wiley & Sons,New York.
September 15, 1993 Mazer/Hendrickson 52
Mazer, S. J. and C"T. Schick. 199 la. Constancy of population and genetic parameters forlife-history and floral traits in Raphanus sativus I. Norms of reaction and the natureofgenotype by environment interactions. Heredity 67: 143-156.
Mazer, S. J. and C. T. Schick. 1991b. Constancy of population and geneticparameters for life-history and floral traits in Raphanus sativus II. Effects ofplanting density on phenotype and heritability estimates. Evolution 45: 1888-1907.
Mazer, S.J., A. A. Snow and M. L. Stanton. 1986. Fertilization dynamics andparental effects upon fruit develo.pment in Raphanus raphanisrrum:consequences for seed size variauon. American Journal of Botany 73: 500-511.
Munz, P. A. and D. D. Keck. 1959. A California Flora. University of CaliforniaPress, Berkeley, CA.
Price, M. V. and N. M. Waser. 1979. Pollen dispersal and optimal outbreeding inDelphinium nelsonii. Nature 277: 294-298.
Rathcke, B. and E. P. Lacey. 1985. Phenological patterns of terrestrial plants.Annual Review of Ecology and Systematics. I6:179 - 214.
Schmitt, J. I983. Individual flowering pbenology, plant size, and reproductivesuccess in Linanthus adrosaceus, a California annual. Oecologia 59:135 -140.
Shreve, F. 1951. Vegetation of the Sonoran Desert. Carnegie Institution ofWashington Publication 591, Washington.
Smith, C. F. 1976. A Flora of the Santa Barbara Region, California. Santa BarbaraMuseum of Natural History, Santa Barbara, CA.
Smith, J. P. and K. Berg. I988. Inventory of Rare and Endangered Vascular Plantsof California. Special Publication No. 1 (4th edition), California Native PlantSociety, Sacramento, CA.
Smith, S. S. E. 1980. Micorrhizas of autotrophic plants. Biological Review 55:475 -510.
Stanton, M. L. I985. Seed size and emergence time within a stand of wild radish(Raphanus raphanistrum L.): the establishment of a fintess hierarchy.Oecologia 67:524 - 531.
Stanton, M. L. and R. W. Preston. 1988. Ecological consequences and phenotypiccorrelates of petal size variation in wild radish, Raphanus sativus(Brassicaceae). American Journal of Botany. 75:528 - 539.
Taylor, D. W. and W. B. Davilla. 1989. Status survey of three plants endemic to theSan Joaquin Valley. Prepared for: Office of Endangered Species, US Fishand Wildlife Service, Sacramento, CA.
September 15, 1993 Mazer/Hendrickson 53
Twisselman, E C. 1967. A'Flora of Kern County California. Wasman Journal ofBiology25:1 - 395.
Young, J. A. and C. G. Young. Collecting, Processing and Germinating Seeds ofWildland Plants. Timber Press, Portland, OR.
We'nt, F. W. 1949. Ecology of desert plants. II. The effect of rain and termperatureon germination and growth. Ecology 30:1 - 13.
54
20"
16"
14"
12"
10 i' ,• _ ] , ,
0 8 _ i - •• diI:i;i
6" :I :dl?_Id_:?:"
4" [-. t ._ --
2" ,_,_i_;±
010 20 30 40 50 60 70
Plant Ht (cm)
Figure la. Frequency distribution for plant height in Foster plot on April 11, 1992.
20"
121
10
o 8U
6"
4" I '
0 10 20 30 40 50 60 70
Plant Ht (cm)
Figure lb. Frequency distribution for plant height in Cuyama population, April 11, 1992.
55
0 i i i i i i i i i i i i i i i i i i i i i
45
40'
35"
30"
25"
151 _KK_:
5 _ _
0 10 20 30 40 50 60 70 80 90 100
Number of fruits per plant
Figure 2a. Frequency distribution for number of developing fruits per plant at Foster Plot on May 2, 1992.
50'
45
40"--
35
30"
2s-
L) 201
15"
,51
00 10 20 30 40 50 60 70 80 90 100
Number of fruits per plant
Figure 2b. Frequency distribution for number of developing fruits per plant at Cuyama population, May 2,1992.
56
45
40 1
35" _
0 25 50 75 ,100 125 150 175 200 225 250
Total number of flowers, developing fruits & buds per plant
Figure 3a. Frequency distribution of total flowers, developing fruits and buds per plant at Foster Plot onApril 11, 1992.
45 I , t , t , I , t , r , t , I t , I , I
40
35
30
= 20
!_ _, _-.
t_
0 25 50 75 100 125 150 175 200 225 250
Total number of flowers, developing fruits & buds per plant
Figure 3b Frequency distribution of total flowers, developing fruits and buds per plant at Cuyama populationon April 11, 1992.
57
4Ill,I,1,1,1,1,111,1,1
12
10 ¸
8̧
4' _}i.
I
0 10 20 30 40 50 60 70 80 90 100
Number of seeds per fruit
Figure 4a. Frequency distribution of number of seeds per fruit at Foster Plot on May 2, 1992.
4 i , i i i , i i i i i i i i i , i i i , i
12]
10] I _.
IoNINI rN0 10 20 30 40 50 60 70 80 90 100
Number of seeds per fruit
Figure 4b. Frequency distribution for number of seeds per fruit at Cuyama population on May 2, 1992.
58
16'
14"
,12'
10'
00 10 20 30 40 50 60 70 80 90 100
Number of ovules per ovary
Figure 5a. Number of ovules per ovary in Foster Plot plants on April 11, 1992.
16"
10-
6-
o 4"
0 10 20 30 40 50 60 70 80 90 100
Number of ovules per ovary
Figure 5b. Number of ovules per ovary in Cuyama population plants on April 11, 1992.
59
70"
60
J
OL)
0 10 20 30 40 50 60 70 80 90 100
Number of open flowers per plant
Figure 6a. Frequency distribution of number of open flowers per plant at Foster Plot on April 11, 1992.
70"
5O
40
o 30
20
10
00 10 20 30 40 50 60 70 80 90 100
Number of open flowers per plant
Figure 6b. Frequency distribution of number of open flowers at Cuyama population on April 11, 1992.
oO
01,111,1,111,111,111,1
60 ¸
50 ¸
40 ¸ _
: 30G_
20
0 10 20 30 40 50 60 70 80 90 100
Number of flower buds per plant
Figure 7a. Frequency distribution of number of flower buds per plant at Foster Plot on April I I, 1992.
80"
5014ot
20 :-'
10 __
0 N_0 5 10 15 20 25 30 35 40 45 50
Number of aborted fruits per plant
Figure 7b. Frequency distribution of number of flower buds per plant at Cuyama population on April II,1992.
61
01*lltltltl*lllllll lit
60"
50"
4O]_
l i_l.Q _
20"_ _
10 .... _
N0 10 20 30 40 50 60 70 80 90 100
Number of flower buds per plant
Figure 8a. Frequency distribution of number of aborted fruits per plant on May 2, 1992 at Foster Plot.
80"
O
30
20
10
00 5 10 15 20 25 30 35 40 45 50
Number ofabortedfruits per plant
Figure 8b. Frequency distribution of number of abo_ed _ui_ per plant at Cuyama population on May 2,1992.
62
Mean ovule number per plant
Error Bars: ± 1 Standard Error(s) ,9O
8O
7O
6O
50
40
3O
20
10
0
Individual Plant Identity
Figure 9. Means and standard errors for ovule numbers in buds from randomly sampled plants (5 buds perplant) at Foster Plot and Cuyama population on April 11, 1992.
63
160 ' ' ' ' y = 13.068- 1.823x + .07x 2 '
140" o/
1201 .
I00" _80" _ o
6o- o "y/40" • _O,_ a__20" O _,_/ " _'O
O__aa o o
0 10 20 30 40 50 60
Plant Ht (cm)
Figure 10. Number of live flowers and buds per plant vs. plant' height in Foster Plot on April I l, 1992 (r2 =.681, p = .00Of).
y = 9.69 - 1.03x + .035x 2100
90 •
80
70 •
60
o_ 50C'4"_ 40' •
•=- 30 Nb •
:_ 20 • •
10' III1•
O' -m.
-100 10 20 30 40 50 60
Plant Ht (cm)
Figure If. Number of fruits on May 2, 1992 at Foster Plot vs. plant height (r 2 : .547, p = .0001)
64
100 " ' ' ' y' =:24"918 + 3.?69x_.O29x 2
90 0 0
80
70 o o o _o60 o oo oo ...9_° _'_0 y/ Oo
°.oo o-
_. 30 .,,"°°'_°o° " oI 20: / o o
10" "0 0
010 15 20 25 30 3; 4o 4; 50 55 60Plant Ht (crn)
Figure 12. Number of seeds per fruit on May 2, 1992 at Foster Plot vs. plant height on April 11 (r_ = .471. p =.0001).
Figure 13. Fecundity (fruit number x seed number per fruit vs. plant height for Foster Plot (r2 = .594, p =.O00l).
65
y = .00 ,2x+ .316, ,r2 =,.0141.2 ' ' ' ' ' '
1 V
V
.8 V_
v "_ vv v
.2 vw V'v .
0 v. v. . vao 30 40 5o 60Plant Ht (cm)
Figure 14. Relationship between plant height on April 11, 1992 and fru!t set on May 2, 1992 at Foster Plot (r2= .014, p = .6389).
y =,-.33,4 + .4,26x • .005,x212 ......
,..;...,.8 • •_• oe8 • _•
J_pw lg _
_D4Z• O0_ •6jw_ --
m 4
2
010 : 20 30 40 50 60
Plant Ht (cm)
Figure 15. Bud length vs. plant height at Foster Plot on April II, 1992 (r2 = .483, p = .0001). At Cuyama therelationship between bud length and plant height is also significant (r2 = .446, p = .0001).
66
y = 1.007 - .275x + .037x2I i i i I • • . , . , .
3 [][] D n []
'2. t [] [] [][] /
[] [] [] _/ []
_ 1.51 0 _ /_ Db
o | •o__nu _[]0- _ -/
03 4 5 6 7 8 9 10 11 12
Bud lengthFigure 16a. Ovary. length regressed on bud length in Foster Plot plants on April 11, 1992 (r2 = .42, p = .0001)
3.5 ' ' Y=''78'5+._91,_-.q33x2i i I i
2.5 []
[] [] DDO
_ 2' [] []m _
_1.5
0 1 [] S [] i
.5 [] [] _ [] []
04 5 6 7 8 9 10 11
Bud length
Figure 16h. Ovary length regressed on bud length at Cuyama population, April I1, 1992 (r 2 --.058, p =.1741).
67
lOG Y ='19"2+9"094x''122x 2 . , .
90' [] []
80'
.-. 70 [] /cq [] [] []60' []
so []
"_ _ []30" m m
20" _ [] []
101 :I []
0 , [].... D,4 5 6 7 8 9 10 11 12
Bud length
Figure 17a. Regression of seed number per fruit on May 2 vs. bud length of next bud to mature on April I1;
Foster Plot plants only. (r2 = .371; p = .0001)
90 ' ' ' ' y '= -35.872 + 11.6,68x : .3,7Yx2, ,
80 a
70 []
_. 50 • [] []
mad= [] [] _.-_u/. 4030 :_ /a :
• [] [] ci: []
10' [] ma mE [] []
04 5 6 7 8 9 10 11
13ud length
Figure'17b. Regression of seed number per fruit on bud length ai Cuyama (r2 = 156, p = .0003)
68
100 ' ' ' Y ='20.023 + 25.1,58x - 3.986x 2 .
901 V v
80"
70" V• v v v
c_ V v V
50_ vVv/...,._.vv Vv
•-_'_'_"40_30 V_-_ "_ v V"v v v
20 v v
10 v
0 .7 v0 .5 1 1.5 2 2.5
Ovary length
Figure lga. Regression of seed number per fruit on ovary length, Foster Plot (r2 = .33, p = .0001).
Figure 25. Frequency distribution of number of developing fruits per plant on May 2, 1993. Both weeded andcontrol plots in all three populations are included.
I 30 .% A ;r•u • % le 15_ 21A _ jL A N".• _ 17 - ';
• .: 1 20 "_"iIS _ B I S P.Q " _ i
• ' • A 39 , p,' "-'1_-_ _" .
,. ..-_._j e __....-_'_'_ A _ •_ J 38 " 8akersfleid
• _'" -- "".*: " :° .... -:_ l_s-_
• _ 0 12%F_ur¢ 2. M_ of the sou,.hem San 27 r- "-,.,
/oaquin VNI_7 d_icting the dL_'ibu¢ion ,_. G "--'Ix
• . I _ . .-_'_ Vcmura . . ,
Map 1. Historical distribution of Caulanthu_ cMifor_icus (from Taylor and Da_illa,1986)
Map 2. Carrizo Plain populations of Caulanthus californicus (from Lewis 1991) andlocations of study sites.
Map 3. Cuyama Valley populations of Caulanthus californicus (from Lewis 1991)and location of Cuyama population study site.
081
Photographs 1 and 2. Caulanthus californicus habitat views. Photo 1 (top):population 4 site (note pink flag in lower left corner). Photo 2 (bottom): Population 4
site looking SW; shrubs are Ephedra californica.
0S2
Photographs 3 and 4. Caulanthus californicus inflorescences. Photo 3 (top):Inflorescence with buds and open flowers. Photo 4 (bottom): open flowers.
083
O84
PhozoBraDh 6. Weeded (right) and con=rol elozs a= po-u!acion 4.
0_5
Appendix C
Field and greenhouse data
Caulanthus californicus Cuyama population: Lab & Field data for 1992
Caulanthus californicus Population 4, Travers Ranch, & Cuyama populations:
Seed number data, 1992
See text for sampling methods.
10ID # population # of seeds (per fruit) mean seed # per fruit
1 1 four 26 25.332 1 four .273 1 four 234 2 four 40 43.675 2 four 676 2 four 24 •7 3 four 41 54.338 3 four 47 •9 3 four 75 °
10 4 four 55 51.0011 4 four 4712 5 four 23 24.6713 5 four 46 •14 5 four: 5= •15 6 four 35 ' 31.6716 6 four 43, °17 6 four 17 I16 7 four I 26 ! 25.0019 7 four 1820 7 four 3121 6 four 15 34,0022 8 four 45 *23 8 four 4224 9 foul 53 50.0025 9 four 5026 9 four 4727 10 four 68 56.6726 10 four 53 •29 10 four 4930 11 four 47 43,0031 11 four 5032 11 four 32 •33 12 four 39 33.6734 12 four 2935 12 four 3336 13 four 55 46.6737 13 four 26 •38 13 four 59 •39 14 four 19 30,6740 14 four 20 •41 14 four 5342 15 four 32 26.6743 15 four 2044 15 four 2845 16 four 21 23.0046 16 four 2047 16 four 2846 17 four 41 37,5049 17 four 3450 16 four 42 42.0051 18 four 42 °52 18 four 4253 19 four 69 72,3354 19 four 7355 19 four 7556 20 four 56 59.0057 20 four 6256 20 four 59
ID # population # of seeds (per fruit) mean seed # per fruit I 0
Caulanthus californicus Population 4: Field data for April 1, 1993
See text for sampling methods.
121quad # treatment pop date whole branches eaten branches # of flwrs # of buds dvlping fruits fruit aborts flwr abort;
1 1 Control Four 4/1 I 0 3 9 5 4 0 " •2 1 Control Four 4/1 1 0 3 2 4 5 0
3 1 Control Four 4/1 2 0 9 13 8 1 04 1 Contro_ Four 4/1 1 0 7 10 9 1 0
5 2 Control Four 4/1 5 1 19 56 3 0 06 2 Control Four 4/1 1 0 6 11 2 0 0
7 2 Contmt Four 4/1 4 ; t 5 43 0 0 08 3 Control Four 4/1 3 _ 2 4 38 0 0 09 3 Control Four 4/1 I = 0 5 9 2 0 0
10 3 Control Four 4/1 4 I 2 8 25 2 0 011 3 Control Four 4/1 4 i 1 13 37 4 0 012 3 Control Four 4/1 1 I 0 5 10 0 0 O13 3 Control Four 4/1 1 i 0 ¢ 8 0 0 014 3 Control Four 4/1 4 ! 2 2 26 4 0 015 3 Control Four 4/1 5 i 4 6 34 0 0 016 3 Control Four 4/1 2 I 0 1 6 0 0 017 3 Control Four 4/1 1 2 : 0 0 0 0 0
18 3 Control Four 4/1 1 0 0 0 0 0 019 3 Control Four 4/1 1 1 _ 0 0 0 0 020 3 Control Four 4/1 1 0 _ 0 0 0 0 021 3 Control Four 4/1 4 1 ! 4 19 0 0 022 3 Control Four 4/1 1 0 0 0 0 0 023 3 Control Four 4/1 1 1 i 0 0 0 0 024 3 Control Four 4/1 1 1 ! 0 0 0 0 025 3 Control Four 4/1 f 0 i 0 0 0 0 026 4 Control Four 4/1 2 0 i 0 3 0 0 027 4 Control Four 4/1 1 0 i 0 2 0 0 028 4 Control Four 4/1 1 0 I 3 8 0 0 029 4. Control FouP 4/1 1 0 I 3 5 0 0 030 4 Control Four 4/1 1 0 I 1 1 0 0 0
31 4 Control Four 4/1 1 0 I 1 5 0 0 032 4 Control Four 4/1 1 0 ! 1 3 0 0 033 4 Control Four 4/1 1 0 I 1 5 0 0 034 4 Control Four 4/1 1 0 I 0 22 0
35 4 Control Four 4/1 1 0 I 4 10 036 4 Control Four 411 2 1 I 0 3 037 4 Control Four 4/1 3 1 I 0 7 038 4 Control Four 4/1 1 0 i 1 5 039 4 Control Four 411 1 0 ! 1 " 5 040 4 Control Four 4/1 1 0 I 0 0 041 4 Control Four 411 1 0 = 0 1 042 4 Control Four 4/1 2 0 ! 0 0 0 "43 4 Control Four 4/1 2 1 , 0 6 044 4 Control Four 4/1 1 1 0 0 0
45 5 Control Four 4/1 1 0 0 6 0 0 046 5 Control Four =4/1 0 1 0 0 0 0 047 5 Control Four i 4/1 1 1 0 3 0 0 0
48 5 Control Four 4/1 1 0 2 8 0 0 049 3 Weeded Four I 4/1 1 0 0 1 0 0 050 3 Weeded Four I 411 1 0 0 0 0 0 0
Caulanthus californicus Population 4: Field data for April 18, 1993
See text for sampling methods.
12Lquad# treatment pop date whole branches eaten branches # of flwm # of buds dvlping fruits fruit aborts f[wraborts
I
1 Con(fol *Fourl 4/18 1 1 'O 0 .... 8 I O2 1 Control Four 4/18 O O 0 S 21 83 1 Control Four 4/18 2 0 0 0 2 4 124 1 Control Four 4/18 O 0 0 0 0 0 05 2 Control Four 4/18 8 2 52 99 28 1 16 2 Control Four 4/18 12 3 38 78 47 2 07 2 Control Four 4/18 5 0 20 34 14 1 0
8 3 Control Four 4/18 4 3 13 29 I 0 0 1
9 3 Control Four 4/18 1 1 2 5 I 0 0 010 3 Control Four 4/18 1 3 O 8 3 3 011 3 Control Four 4/18 1 4 0 12 0 0 012 3 Control Four 4/18 5 4 28 39 19 0 013 3 Control Four 4/18 3 6, 17 22 11 0 014 3 Control Four 4118 I 0 : 5 11 1 0 0
15 3 Control Four 4/18 1 1 i 7 9 0 0 016 3 Control Four 4118 2 0 0 6 2 0 017 3 Control Four 4118 1 0 0 5 0 0 218 3 Control Four 4118 2 1 2 12 0 0 019 3 Control Four 4/18 0 1 0 0 0 0 020 3 Control Four I 4/18 1 0 1 , 6 5 3 2
21 3 Control Four, 4/18 3 1 21 I 20 30 0 0
22 4 Control Four 4/18 2 5 0 I 11 0 0 023 4 Control Four 4/18 1 0 2 2 1 0 0
24 4 Control Four 4/18 1 1 O 0 0 0 025 4 Control Four 4/18 1 0 0 0 0 0 026 4 Control i Four 4/18 1 0 0 0 0 0 0
27 4 Control Four 4/18 1 0 0 3 1 0 0281 4 Control Four 4/18 1 O 0 2 1 0 4291 4 Control Four 4/18 1 0 0 4 2 0 330 4 Control Four 4/18 2 1 0 2 0 0 031 4 Control Four 4/18 1 0 0 4 8 0 032 4 Control Four 4/18 0 3 0 0 0 0 O33 4 Control Four 4/18 2 1 0 2 0 0 034 4 Control Four 4118 0 1 0 0 0 0 035 4, Control Four 4/18 1"i 0 0 2 0 0 036 4 Control Four 4/18 1 i 1 0 0 0 0 0
37 4 i Control Four 4/18 2 I 2 0 7 0 0 038 4 I Control Four 4/18 1 0 2 4 0 0 0
39 4 ! Control Four- 4/18 0 2 0 0 0 0 .040 4 I Control Four 4/18 0 1 0 0 0 0 041 4 Control Four 4/18 1 1 1 1 0 0 0
42 4 Control Four 4/18 1 3 0 2 0 0 043 4 Control Four 4/18 1 0 0 4 0 _ 0 044 5 Control Four 4/18 1 0 1 0 0 = 0 0
45 5 Control Four 4/18 0 0 0 0 0 I 0 0
46 5 Control Four 4/18 0 0 0 0 0 I 0 O47 1 Weeded Four 4t18 1 2 7 0 0 I 0 048 3 Weeded Four 4/18 1 0 0 40I 0 0
49 3 Weeded Four 4/18 0 2 0 0 0 I 0 050 3 Weeded Four 4/18 I 1 0 5 0 I O 0
51 3 Weeded Four 4/18 I 0 0 2 0 I .0 252 3 Weeded Four 4/18 I 0 0 2 0 0 053 3 Weeded Four 4/18 1 0 1 1 0 0 054 3 Weeded Four 4/18 1 3 0 7 O 0 0
55 3 Weeded Four 4/18 1 1 0 10 0 0 056 3 Weeded Four 4/18 0 3 0 0 0 0 057 3 Weeded Four 4118 1 1 2 9 0 O 0
58 3 Weeded Four 4/18 1 3 0 8 "0 0 059 3 Weeded Four 4/18 0 0 0 0 0 0 O60 3 Weeded Four 4/18 3 5 0 20 0 0 061 3 Weeded Four 4118 1 1 0 0 0 O O
62 3 Weeded Four 4118 1 0 0 2 0 0 0
125_quad# treatment pop date whole branches eaten branches # of flwrs # of buds dvrpingfruits fruit aborts flwraborts
63 3 Weeded Four 4/18 1 4 O 8 0 0 064 3 Weeded Four 4/18 1 3 0 7 0 0 068 3 Weeded Four 4/18 1 2 0 3 O 1 066 3 Weede_ Four 4/18 5 14 0 39 0 0 0
87 3 Weeded Four 4/18 5 15 O 25 0 0 168 3 Weeded Four 4/18 1 1 0 7 0 0 0
71 3 Weeded Four 4/18 0 O 0 0 O 0 072 3 Weeded Four 4/18 5 14 14 18 5 0 07:3 4 Weeded Four 4/18 I O 1 2 O 0 074 4 Weeded Four 4/18 I 0 O 1 I 0 O 0
75 4 Weeded Four 4/18 1 1 0 3 = 0 O O78 4 Weeded Four 4/18 1 0 O 5 0 0 077 4 Weeded Four 4/18 1 1 2 0 : 0 0 078 4 Weeded Four 4/18 1 2 0 2 0 0 0
79 4 Weeded Four 4/18 - 1 2 O 7 0 0 080 4 Weeded Four 4/18 1 O 0 0 0 0 5
81 4 Weeded Four 4/18 1 O I O 0 0 0 282 5 Weeded Four 4/18 1 0 i O 5 0 0 083 5 Weeded Four 4/18 5 5 4 97 0 0 0
8, , weededFour4/18 ol OI OI 8 o 0 o85 5 Weeded ,Four 4/18 0 I 0 I 0 I 0 0 0 0
126
Caulanthus californicus Population 4: Field data for May 2, 1993
See text for sampling methods.
].2quad# treatment pop date whole branches eaten branches # of flwrs # of buds dvlping fruits fruit aborts flwraborts
1 1 Control Four 512 2 0 0 0 0 • 0 02 1 Control Four 5/2 1 0 0 0 0 0 03 2 Control Four 5/2 0 3 0 0 0 0 04 2 Control Four 5/2 2 1 0 0 4 9 16
5 2 Control Four 5/2 1 4 0 0 14 12 176 3 Control Four 5/2 0 2 0 0 0 0 07 3 Control Four 5/2 2 4 6 0 6 4 148 3 Control Four 5/2 0 4 0 0 0 0 09 3 Control Four 5/2 1 4 0 7 t 0 O.
10 3 Control Four 512 1 0 0 0 1 0 0
t 1 3 Control Four 5/2 1 1 1 0 0 0 0t 2 3 Control Four 5/2 0 0 0 0 0 0 013 3 Control Four 5/2 0 0 0 0 0 0 014 3 Contmt Four 512 0 0 0 0 0 0 0
15 3 Control Four 5/2 0 0 0 0 0 0 016 3 Controt Four 5/2 0 0 0 0 0 0 017 3 Control Four 5/2 0 0 0 0 0 0 0
18 3 Control Four 5/2 0 0 0 0 0 0 019 4 Control Four 5/2 3 9 0 16 0 0 020 4 Control Four 5/2 2 0 0 0 0 0 021 4 Control Four 5/2 0 2 0 0 0 0 0
22 4 Control Four 5/2 1 2 0 2 0 0 023 4 Control Four 5/2 2 I 4 8 0 0 124 4 Control l Four 5/2 0 4 1 0 0 0 025 4 Contml i Four 5/2 1 8 1 5 0 0 026 1 Weeded _ Four 5/2 1 2 3 4 3 0 027 3 Weeded ! Four 5/2 2 3 0 12 0 0 028 3 Weeded = Four 5/2 0 2 0 0 0 0 029 3 Weeded, Four 512 1 0 0 1 0 0 030 3 Weeded Four 5t2 3 9 0 13 0 0 031 3 Weeded Four 5/2 0 2 0 0 0 0 032 3 Weeded Four 5/2 I 4 0 6 0 0 0
33 3 Weeded Four 5/2 1 3 I 0 4 0 0 034 3 Weeded Four 5/2 0 5 I 0 0 0 0 035 3 Weeded Four 5/2 0 31 0 0 0 0 0
36 3 Weeded Four 5/2 0 0 I 0 0 I 0 0 037 3 Weeded Four 5/2 0 0 I 0 0 I 0 0 0
38 3 Weeded Four 5/2 0 0 I 0 0 I 0 0 039 3 Weeded Four 5/2 0 0 I 0 0 I 0 0 0
40 3 Weeded Four 5/2 0 0 I 0 0 I 0 0 I 041 3 Weeded Four 5/2 0 0 I 0 0 I 0 0 I 042 3 Weeded Four 5/2 0 0 = 0 0 i 0 0 I 043 3 Weeded Four 5/2 0 0 0 0 0 I 0 044 3 Weeded Four 512 0 4 0 0 0 I 0 045 3 Weeded Four 5/2 0 0 0 0 0 ' 0 046 3 Weeded Four 5/2 0 0 0 0 0 0 047 3 Weeded Four 5/2 0 0 0 0 0 0 048 3 Weeded Four 5/2 0 0 0 0 0 0 0
59 6 Weeded Six 5/2 ' 1 1 0 0 2 2 560 6 Weeded S x I 5/2 3 4 8 20 57 2 8
61 7 Weeded Six I 5/2 1 4 0 4 0 0 062 7 Weeded S x I 5/2 1 1 0 0 2 0 0
63 7 Weeded Six 512 t I 0 0 3 0 0
64 7 Weeded Six 512 1 4 0 0 1 0 0
65 8 Weeded Six 5/2 6 5 10 22 45 22 4
138
Caulanthus ca!ifornicus Foster Plot, Population 4, & Population 6:
Field survivorship data for 1993
See text for sampling methods.
139
140
141
Caulanthus californicus Within and Between site Greenhouse crosses, 1993
See text for explanation.
142.
1,13
144
145
146
14"/
UNIVERSITY OF CALIFORNIA, SANTA BARBARA
BERKELEY • DAVIS * [RVINE • LOS ANGELES • RIVERSIDE • SAN DIEGO • SAN FRANCISCO _/ SANTA BARRARA SANTA CRUZ
DEPARTMENT OF BIOLOGICAL SCIENCES SANTA BARBARA, CALIFORNIA 93106-9610PHONE: (805) 893-3511
FAX: (805) 8934724
Mazer: (805)-893-8011 £_ _. L
17, 1993Gmail: mazer @li fesci.uesb.edu aeptemuer
Ms. Sandy MoreyNatural Heritage DivisionCalifornia State Department of Fish and Game1.416 Ninth Street
Sacramento, CA 95814
Dear Sandy,
Enclosed you will find the final report describing our field and laboratorywork, results, and recommendations concerning Caulanthus californicus . Thisreport fulfills our commitment to Contract FG 1461.
If you have any questions concerning details of the methods or resultsdescribed in this or in our previous reports on woolly threads and the kern mallow,please don't hesitate to call or FAX me at the phone numbers listed above. I amsending a copy of this report to Rodd Goss and to Ivana No61, a new District Botanistof the BLM. She recently graduated with a Master's degree from UCSB and is nowstationed at the BLM office in Bakersfield.
Let's keep in touch,
Sincerely,
Susan J. Mazer
Associate Professor of Biology
UNIVERSITY OF CALIFORNIA, SANTA BARBARA
BERKELEY * DAVIS • IRVINE * LOS ANGELES * RIVERSIDE * SAN DIEGO * SAN FRANCISCO _I SANTABARBARA SANTACRUZ
DEPARTMENT OF BIOLOGICAL SCIENCES SANTA BARBARA, CALIFORNIA 93106-9610PHONE: (805) 893-3511
Ms. Sandy MoreyNatural Heritage DivisionCalifornia State Department of Fish and Game1416 Ninth StreetSacramento, CA 95814
Dear Sandy,
Enclosed is an unbound copy of the Caulanthus californicus report. Just soyou know, I have no funds available for this xeroxing (nor any help to do it), so allthe time and copying charges come out of my pocket!!