B. Current Taxonomic Status
History: How many manzanita species have existed within San Francisco is difficult to
determine. At the time of early botanical explorations of San Francisco in the 1800s, there were
very few described species of Arctostaphylos and the diversity of the genus would not be
recognized or understood for some time. Most of the early collections, consequently, were
assigned names that were already available, A. pungens (Mexican manzanita) and A. pumila
(sand-mat manzanita) (e.g., Behr 1888, 1891, Brandegee 1892). Based on his memory, Behr
(1891) described much of San Francisco as having chaparral dominating most of the hilltops and
ridgetops in San Francisco, principally by Ceanothus thyrsiflorus (blue blossom), but also with
large numbers of manzanitas (Arctostaphylos species).
Eastwood (1905) was the first to describe a San Francisco manzanita as a new species. She
recognized that the prostrate species on serpentine at the Laurel Hill Cemetery were not A.
pumila. She noticed it was distinguished from A. pumila by several characters and named it A.
franciscana (Franciscan manzanita). The type locality was Laurel Hill Cemetery (now Laurel
Hill Playground in Laurel Heights). She correctly interpreted A. pumila as a Monterey Bay
endemic. The low spreading Franciscan manzanita at the Laurel Hill Cemetery was already just
a remnant of manzanitas that were previously known, including erect individuals of perhaps now
extirpated species (Behr 1891).
Arctostaphylos montana subsp. ravenii (The Presidio manzanita) was probably discovered quite
early, during an early listing of plant species of San Francisco, but was called A. pungens at that
time (Brandegee 1892, Raven 1952). In a San Francisco flora, under A. pungens, Brandegee
(1892) refers to “Mr. C. A. Michener collected it last year somewhere between Lobos Creek and
Fort Point.” Peter Raven was the first to rediscover the Presidio manzanita (Raven 1952); he
referred to the Brandegee article and thought they had probably collected from the same
individual. Instead of A. pungens, Raven interpreted the plant as A. franciscana, noting that it
was a good match for a collection from Mt. Davidson. A flora of San Francisco was published a
few years later also interpreting the plant as A. franciscana (Howell et al. 1958). They describe
the plant as historically having a distribution of north of Baker’s Beach, Laurel Hill Cemetery,
Masonic Cemetery, and Mt. Davidson. Later, Wells (1968) determined that this plant had a
different ploidy level (number of sets of chromosomes) from the Franciscan manzanita (A.
franciscana). Because the Presidio manzanita also differed in some morphological
characteristics, Wells (1968) separated it from the Franciscan manzanita and described it as a
subspecies of a Monterey Bay species, and used Raven’s name as an honorific, A. hookeri subsp.
The recent discovery of A. franciscana along Doyle Drive in a serpentine outcrop (visible in
historic aerials prior to the construction of Doyle Drive) suggests they may have co-occurred in
the Presidio. Leaf sizes and morphology are variable, but A. franciscana generally is roughly
twice as long as it is wide for all the herbarium samples at the California Academy of Sciences,
the UC and Jepson Herbaria (ratios general close to 0.5, but ranging from 0.42-0.66 for all
leaves), while the Presidio manzanita is a rounder leaf (with width:length ratios ranging from
0.62 to 1.0). The range of variation in the Franciscan manzanita may still overlap that of the
Presidio Manzanita depending on how the Mt. Davidson collections are interpreted. If some are
sorted out as representing the Presidio manzanita based on leaf shape, then the Presidio
manzanita may have had a much greater range than the serpentine soils of the Presidio and co-
occurred with the Franciscan manzanita there as well.
The history of San Francisco manzanitas is not quite over. Recently, an additional manzanita
was found in the Laurel Hill area, in the yard that was formerly that of the Laurel Hill Cemetery
caretaker. This manzanita was determined to be a new subspecies of A. bakeri (a Sonoma
County serpentine endemic) (V.T. Parker, unpubl. data).
Taxonomic status, 1968-1999: When Wells (1968) named the Presidio manzanita, he also
interpreted a cluster of coastal species as being close relatives and made them all subspecies of
A. hookeri. Wells felt the five taxa shared sufficient morphological similarities that they were all
just variants of each other. Along with the nominate species, Wells also placed A. montana, A.
franciscana, and A. hearstiorum as subspecies of A. hookeri, in addition to naming A. hookeri
subsp. ravenii. Arctostaphylos hookeri has shiny green leaves and small fruit like the Presidio
manzanita, but differs in other ways. The three northern subspecies are all serpentine endemics,
for example. Two of the three northern taxa (subsp. ravenii and subsp. montana) are tetraploid
species with twice the number of chromosomes. The other northern taxon (subsp. franciscana)
and the two southern subspecies are diploids. Interpreting this cluster of five taxa as all
subspecies of a single species resulted in an A. hookeri complex of widespread geographic
Arctostaphylos hookeri subsp. hearstiorum is known only from the coastal prairie and chaparral
areas from the Hearst Ranch near San Simeon, San Luis Obispo County. Arctostaphylos hookeri
subsp. hookeri is restricted to chaparral areas near the coast on sands and clays in the vicinity of
the City of Monterey; the northernmost population is just across the county boundary into Santa
Cruz County on both sides of Hwy 1, south to the Carmel area, and west only a few kilometers
into Ft. Ord and on the slopes of the county park at Jack’s Peak. Arctostaphylos hookeri subsp.
francicana and A. hookeri subsp. ravenii are known only from serpentines and greenstones
within the county of San Francisco; the former is extirpated from the wild and the latter
represented only by the single known individual. Arctostaphylos hookeri subsp. montana is
endemic to Marin County and is found almost entirely on serpentine soils (principally within the
Mt. Tamalpais watershed of the Marin Municipal Water District); its populations are large and
extensive within the serpentine areas on the watershed.
Wells later maintained the Presidio manzanita as a subspecies of A. hookeri in his treatment of
Arctostaphylos for the first edition of the Jepson Manual: Higher Plants of California (Hickman
1993) and in a self-published monograph on Arctostaphylos (Wells 1993, 2000).
Current taxonomic status: Markos et al. (1999) re-evaluated the Wells (1968, 1993, 2000)
treatment of the A. hookeri complex. They used genetic sequences to determine relationships
and included a number of other manzanita species in their analyses. They found that the two
southern subspecies were related closely by nuclear ribosomal spacer DNA sequences, but that
the 3 northern subspecies differed considerably from the 2 southern subspecies. Subsequent
genetic work using a variety of nuclear genes has confirmed the distinction of the northern
diploid species, A. franciscana (Franciscan manzanita), from the two southern diploid
subspecies. The two tetraploid taxa are more complex because they contain twice the number of
genetic alleles. Their genetics suggests that the tetraploids (subsp. ravenii and subsp. montana)
may be the result of complex hybridization among diploids followed by polyploidization to
increase fertility. Each contains alleles in common with A. franciscana, suggesting the
Franciscan manzanita may have been a parent historically, and subsp. ravenii has alleles in one
of the genes that suggests A. hookeri subsp. hookeri may have been one of the historic parents in
its lineage (Appendix A).
Because of these and other subsequent analyses, the taxonomic status of the A. hookeri complex
has been reevaluated. Currently, the Presidio manzanita has been placed as a subspecies of
Arctostaphylos montana (A. montana subsp. ravenii, Parker et al. 2007). The Presidio manzanita
is found as A. montana subsp. ravenii in the Flora of North America (Parker et al. 2009) and the
second edition of the Jepson Manual (Parker et al. 2010).
C. Evolution of Arctostaphylos: Where does the Presidio Manzanita fit in?
Arctostaphylos Adans. (Ericaceae) is composed of over 100 taxa and is the most species-rich
genus of subfamily Arbutoideae Nied. (Wells 2000, Parker et al. 2009). Except for the
circumboreal Arctostaphylos uva-ursi (L.) Spreng. these species are confined to western North
America and Mexico. In California, species of Arctostaphylos occur in a variety of plant
communities, including chaparral, montane scrub, mixed-coniferous forests, serpentine barrens,
and oak savannas. At low latitudes, Arctostaphylos uva-ursi occurs in montane regions (e.g.,
Rockies, Appalachians, Alps, Himalayas) and in coastal pine barrens and dunes (e.g., New
Jersey and Michigan, USA). A disjunct population of A. uva-ursi (called A. uva-ursi subsp.
cratericola (Donn.Sm.) P.V.Wells at one time) in the volcanic highlands of Guatemala
represents an odd outlying distribution.
Arctostaphylos is a part of the Ericaceae subfamily Arbutoideae with five other genera: Arctous
Nied., Arbutus L., Comarostaphylis Zucc., Ornithostaphylos Small, and Xylococcus Nutt. (Diggs
& Breckon 1981). Molecular and morphological studies of relationships (phylogenies) in the
Ericaceae place the Arbutoideae at the base of the family tree with the Monotropoideae (a
subfamily of mycotrophic parasitic plants like pine drops and the snow plant) sister to the rest of
the family (Kron et al. 2002). Phylogenetic trees for 16 species of Arctostaphylos were
developed using molecular data from the nuclear ribosome that yielded a structure in which
species were found in one of two distinct clades (a clade is a complete set of branches; thus,
Arctostaphylos would have 2 sets of branches or lineages if it has a 2 clade topology) (Markos et
al. 1998). The two-clade structure recovered in their trees differed considerably from the
concepts of Wells (2000).
The evolutionary structure of the two clades also did not support the monophyly of the A.
hookeri species complex in Wells’ section Pungens (Markos et al. 1998) (monophylly refers to
an ancestor and all its descendants without any exception). Boykin et al. (2005) presented an
additional phylogeny containing 38 taxa that also recovered a two-clade topology. Further, these
two-clade phylogenies are congruent (with respect to species composition) with a phenogram
constructed from flavonoid data for 20 species of Arctostaphylos (Denford 1981). This leads to
the interpretation that Arctostaphylos has two lineages; one clade represents a cluster of species
that are much more closely related to one another than to species in the other clade. Because
fossils of Arctostaphylos are found in the mid-Miocene (over 15 MYBP, million years before
present), these lineages may represent a very old divergence in the genus. The most current data
support the 2 clade interpretation (Wahlert et al. 2009).
The center of diversity for Arctostaphylos is in the California Floristic Province, principally from
the San Francisco Bay area south to Santa Barbara County. The coastal areas and Coast Ranges
of California, from Mendocino County to Santa Barbara County, are particularly rich in taxa
(Markos et al. 1998). Arctostaphylos has a high percentage of endemic taxa, many of which are
edaphic specialists (Wells 2000). Nearly half of all California taxa are considered rare,
threatened, or endangered (CNPS 2009). The diversity of modern taxa may be the result of a
radiation of Arctostaphylos since the end of the Miocene (5.3 MYBP) (Edwards 2004). Factors
contributing to the diversification of the genus include local adaptation to California’s diverse
substrates, an increase in fire frequency at the end of the Pliocene, the onset of a Mediterranean
climate, and the contemporaneous expansion of chaparral into forested areas (Raven & Axelrod
1978; Axelrod 1989; Edwards 2004).
Hybridization and polyploidization are two processes driving speciation in Arctostaphylos. The
base chromosome number in Arctostaphylos is x = 13, with most species being either diploid (2n
= 26) or tetraploid (2n = 52). Many taxa have been proposed as arising via a process called
recombinant hybrid speciation. Recombinant hybrid species means hybridization occurs
between two different diploid species and their hybrid progeny recombine genetic combinations
until one set stabilizes successfully (e.g., Dobzhansky 1953; Gottlieb 1968; Schmid et al. 1968;
Kruckeberg 1977; Knight 1985; Keeley & Massihi 1994; Parker & Vasey 2004). For a group of
46 edaphic specialists or endemic taxa, for example, Wells (2000) proposed a hybrid origin for
Two processes can yield polyploids such as the tetraploids found in Arctostaphylos. One process
is called autopolyploidy, and involves a single species. Differing ploidy levels observed in
Arctostaphylos uva-ursi (2n = 26, 39, 52 and 78) suggests recent and recurring polyploidization
by autopolyploidy, which is common in arctic plants (Rosatti 1981; Brochmann et al. 2004).
Another process, allopolyploidization, is also potentially common in Arctostaphylos. In this
process, two different diploid species hybridize, but their hybrid progeny have low fertility;
under these conditions, sexual gametes often fail to ‘reduce’ the number of chromosomes and
this leads to polyploids. In this case, the polyploids have genetics from parents of two different
species. Schierenbeck et al. (1992) provide morphological and cytological evidence for the
allopolyploid origin of the tetraploid A. mewukka arising from crosses between the diploids A.
patula and A. viscida.
The Presidio manzanita is a tetraploid species in Arctostaphylos. The origin of tetraploids like
the Presidio manzanita is not clear, but some nuclear genes suggest relationships to species from
both clades including A. hookeri, A. franciscana, A. ohloneana, A. pungens, and several other
potential species (Parker, unpublished data, Appendix 4). There is considerable overlap, but not
complete overlap, with that of A. montana, a species that is similar in ecology and morphology to
the Presidio manzanita.
A. Habitat Type
Arctostaphylos montana subsp. ravenii is found as an individual in a maritime chaparral-coastal
prairie overlooking the Pacifica Ocean at the margin of remnant maritime chaparral (Figure 1).
Interpretations of older collections (Table 2) suggest it should also be considered a chaparral
species. Edaphically, it is distributed on serpentine soil, which are quite poor in plant resources,
like nitrogen and phosphorus, have an imbalance in proportions of available calcium and
magnesium, and usually contain toxic levels (for most plants) of some heavy metals. Grassland
serpentine soils are generally less extreme in these characteristics than thinner soils derived from
serpentine. The vegetation with which it is found is characterized by nutrient poor soil, foggy
moderated climates, and wildfire.
Because of the moderating influence of coastal fog, the Presidio manzanita would be considered
a member of maritime chaparral. Maritime chaparral represents a plant community of special
concern because of the high density of narrowly distributed endemic species (the California
Coastal Commission, for example, considers it an Environmentally Sensitive Habitat). From
Mendocino County to the Channel Islands, this chaparral community is patchily distributed
associated with forests edges or odd soils.
Local chaparral communities along the central California coast demonstrate unique combinations
of species. Around Lake Merced, two manzanitas are still found, quite isolated and endangered,
along Brotherhood Way. This is the type locality for A. crustacea subsp. rosei; this species was
originally thought to be another San Francisco endemic, but later populations were found along
the Big Sur coastline. Individuals of A. crustacea subsp. crustacea are also found around Lake
Merced (now restricted to Brotherhood Way). To the south of San Francisco, chaparral on
Montara Mt. is dominated by two maritime-restricted manzanitas (Arctostaphylos montaraensis
and A. crustacea), one found only on Montara Mt. and San Bruno Mt. San Bruno Mt. contains
five manzanitas, three found only on the mountain. In the East Bay, remnant sites harbor an
endemic of those hills along with a more widespread manzanita (A. pallida and A. crustacea).
North of San Francisco, along western Bolinas Ridge and the western and southern slopes of Mt.
Tamalpais, for example, 3 species of Arctostaphylos (A. glandulosa subsp. glandulosa, A.
virgata, and A. sensitiva) are common, while Inverness Ridge contains a mostly different
collection (A. uva-ursi, A. glandulosa subsp. cushingiana, A. virgata and A. x repens). Given
this local endemism within the fog zone, the diversity found within San Francisco is not
Forty-three of the 62 species of Arctostaphylos globally are found in the maritime zone of the
California central coast. Arctostaphylos is characterized by a large number of taxa (species or
subspecies) that are highly restricted in distribution. Globally, 66 species and subspecies (out of
96 total species and subspecies) are local endemic, highly restricted taxa. Of all these local,
highly restricted taxa, 50 of them (73%) are restricted to the maritime zone. Around one third of
Ceanothus taxa globally are also restricted to maritime habitats including the common blue
blossom of the Presidio (C. thyrsiflorus). Most of these shrubs listed as CNPS 1B.1, State or
Federally endangered status, and they tend to co-occur with with a number of herbaceous species
with a similar status ranking. Other areas of maritime chaparral exhibit similar patterns of high
diversity with a high proportion of rare and endemic plants. The Monterey Bay region, for
example, contains 12 local endemics in Arctostaphylos, including Arctostaphylos glutinosa, A.
pajaroensis, and A. montereyensis. In addition to these 12 taxa, other species restricted to
maritime chaparral but more widely distributed are also found, including A. sensitiva and A.
crustacea subsp crustacea.
Information about maritime chaparral in the Monterey region can be found in Griffin (1978),
Greenlee and Langenheim (1990), Holl et al. (2000), Van Dyke et al. (2001), Parker and Vasey
(2005) and Vasey and Parker (2008). The San Luis Obispo region south to northern Santa
Barbara County also has similar high endemism in chaparral restricted to maritime influences
(Wells 1962; Davis et al. 1988, 1989, Tyler 1996, Tyler and Odion 1996, Odion 2000, Odion and
Davis 2000, Odion and Tyler 2002).
Maritime fog climate:
The Presidio is characterized by summer fog with relatively cool temperatures. While official
temperatures are kept for San Francisco in the Mission and at the airport, the outer coast is
usually much cooler. Monterey, for example, has a seasonal average difference between average
winter and summer temperatures of only 3.9 degrees C (7 degrees F) (NOAA). Pt. Reyes has
similar average summer and winter temperatures with a seasonal range of less of 3 degrees C (5
deg. F) (Howell 1970).
Figure 8. Influence of coastal fog is critical for species occupying communities in the maritime zone. Temperature
and moisture ranges are highly modified and enable very different groups of species to survive. Photo by V.T. Parker.
More interior sites even close by could experience a much greater seasonal range of about 11
degrees C (20 degrees F). This maritime climate is an important aspect of these coastal plant
communities (Griffin 1978). Howell (1970) comments that Marin County has two distinct life
zones (after Merriam) defined by the coastal maritime climate and the interior climate. From
Mendocino to Santa Barbara Cos., the immediate coastline is impacted to a greater or lesser
degree by fog, and creates a transitional area between the Pacific Coast maritime climate (of
coastal Washington to northern California) and the southern California or interior Mediterranean
climate; in the transitional area, fog moderates the summer-dry stressful conditions (Azevedo
and Morgan 1974, Kreissman 1991, Dallman 1998) (Figure 7).
Serpentine grasslands and chaparral:
The Presidio manzanita additionally is an endemic to serpentine soils. Serpentine soils are
highly infertile (low calcium: magnesium ratio, low nitrogen content, high heavy metal content,
and low water holding capacity). As a consequence, these odd soils contain a large number of
endemics (Kruckeberg 1984, Huenneke et al. 1990, Safford et al. 2005). Many introduced
species are not successful on serpentine soils in most of California; serpentine sites, therefore,
tend to have a greater proportion of natives than similar sites on other soils in adjacent areas.
Serpentine areas are well known for their high endemism (Kruckeberg 1984, Safford et al. 2005).
Within Arctostaphylos, eight taxa are wholly or nearly restricted to serpentine (Safford et al.
2005, Parker et al. 2009). For the California flora as a whole, almost 13% of endemics are
restricted to serpentine soils; 15% of all plant taxa listed as threatened or endangered are
serpentine endemics or have some degree of association with ultramafic substrates (Safford et al.
2005). Serpentine chaparral and grasslands are similar in these patterns. For more information,
see Huenneke et al. 1990, Harrison 1999, Harrison and Viers 2006, Harrison et al. 2006b.
Because there are non-native species in the grassland area containing the Presidio manzanita,
some concern should be considered in their overall impact. The dynamics of serpentine areas are
generally dependent on their poor nutrient status and summer water stress. Species that may
modify either of these processes might tip the balance toward a conversion of the site.
Introduced species create a different underground ecosystem for soil biota, due to the types of
chemicals they secrete, the chemistry of their roots and their density, and the amount of carbon
lost out of roots. As a consequence, plants create different microbiotic ecosystems by their
influences (Grayston et al. 1996) and invasives may increase their success by inhibiting or
shifting composition of soil mutualists like mycorrhizal fungi or beneficial bacteria (Callaway et
Arctostaphylos species, for example, are dependent on a mutualistic relationship with
mycorrhizal fungi for their nutrition in the poor soils they tend to occupy (Horton et al. 1999).
They form arbutoid mycorrhizae, a variation on ectomycorrhizae. Other species within the
grassland they occupy form arbuscular mycorrhizae (older name is VAM or endomycorrhizae).
These types of species do not overlap in fungal species with manzanitas in their mutualisms
(Horton et al. 1999). Invasion by non-native species can shift the soil microbial biota (Grayston
et al. 1996, Callaway et al. 2004, Batten et al. 2006) and serpentine soils are subject to some
invasion with impacts on diversity (Huenneke et al. 1990, Harrison et al. 2006a).
V. Fire Ecology and Population Dynamics
A. Response to Fire
Maritime chaparral and adjacent coastal communities may contain a large number of rare species
and be confined to foggy, cooler conditions for California, but the dynamics of the vegetation are
driven by a fire regime dominated by canopy wildfire like other chaparral and coastal prairie-
coast scrub regions. Maritime chaparral often is associated with closed-cone species that are
restricted to the immediate coast, like Pinus muricata, P. radiata, P. attenuata, Hesperocyparis
(Cupressus, see Adams et al. 2009) abramsiana, H. goveniana, H. macrocarpa and the
resprouting trees of the mixed evergreen forest. Classic coastal maritime communities like
redwood forests are clearly dependent on fire and Sequoia sempervirens (coast redwood) can
resprout after fire damage and depends of fire for seedling establishment.
A diversity of life history responses occurs among all the species that are a part of maritime
chaparral (e.g., Keeley 2000, Keeley and Zedler 1978, Keeley et al. 2006, Parker 1987a, b, 1990,
1993, Parker and Kelly 1989). The principal life history types illustrated in these studies
1) annual plants dependent on persistent dormant soil seed banks; these species flourish
the first year or two and rebuild a seed bank. Without cues from a fire, their seeds remain
dormant thereafter. Examples are many and include species within Phacelia and Dicentra.
2) herbaceous perennials; while usually present every year to some extent, productivity
and flowering are often restricted to the years following wildfire. This is the time seedlings
establish and new individuals get started. Example genera include Calochortus and Zigadenus;
3) obligate sprouting woody species; these species establish new individuals in old
growth chaparral. They depend on resprouting to persist in wildfire areas. Example genera
include Quercus, Rhamnus, and Heteromeles.
4) facultative sprouting woody species are those that can resprout after fire but also
produce persistent soil seed banks (also called facultative seeders by some). Their seeds
germinate the first year after fire, establishing new individuals. Three genera in California have
species that exhibit this life history, Arctostaphylos, Ceanothus and Adenostoma. Other genera
have some species that are part of coastal scrub and often overlap into chaparral, especially
immediately post-fire. These species may resprout in areas with lower fire intensities and do
have a portion of their seeds within a soil seed bank; example genera are Salvia or Lepechinia.
5) obligate seeding woody species are those in which fire kills the adult individuals. The
recovery of the population is dependent on seed germination and the establishment of new
individuals from a persistent seed bank. Example shrub genera in California are Arctostaphylos
and Ceanothus; Pinus and Hesperocyparis (Cupressus) also have species with closed cones that
function the same.
Three of these life history types depend on fire for stimulation of their seed banks for
reproduction: the annuals, the facultative sprouting woody species, and the obligate seeding
woody species. The herbaceous perennial species do not seem to do well except in the post-fire
environment although they do not produce seed until the end of the first post-fire growing
season, while the obligate sprouting woody species reproduce in older stands of chaparral (see
Parker and Kelly 1989, Keeley 2000). When summarized by important characteristics (Table 1),
all the life history types depend on some aspect of the fire regime, but not all of them contain the
same characteristics. Because the Presidio manzanita and its relatives within Arctostaphylos
only exhibit two of these responses, those life history types will be examined in more depth.
Two life history types:
Arctostaphylos and Ceanothus each contain two types of species in terms of their life histories:
one group of species that contain burls or root crowns with dormant buds that permit resprouting
after fire (facultative sprouters), and another that are unable to resprout after fire or cutting and
are killed by fire (obligate seeders) (Keeley and Zedler 1978, Keeley 1991, Parker and Kelly
1989). The name ‘facultative sprouter’ means that they tend to survive fire and sprout new
aboveground growth quickly while also having seeds in a dormant persistent soil seed bank that
responds to a stimulus from a fire. Obligate seeders refer to the fact that adults are killed by fire
Table 1. Summary of different life history characteristics found among chaparral plants. Arctostaphylos species fall
into two of these five categories (facultative sprouters or obligate seeders).
and populations only reestablish from persistent soil seed banks. Facultative sprouters resprout
after fire or cutting that also have persistent soil seed banks; these plants are sometimes called
facultative seeders. This life history type is not as common within Arctostaphylos or Ceanothus,
but facultative seeders are often dominant species. In Marin County, only Arctostaphylos
glandulosa (subsp. cushingiana on Pt. Reyes, subsp. glandulosa common elsewhere), some of
the A. uva-ursi populations, and Adenostoma fasciculatum (chamise) illustrate this life history
type. In the southern part of San Francisco, and within the Santa Cruz Mts to the south of San
Francisco, are Arctostaphylos crustacea (subsp. crustacea and rosei in San Francisco, subsp.
crustacea and subsp. crinita in the Santa Cruz Mts).
Shrubs and trees killed by fire and that depend on dormant seeds banks are the obligate seeders.
This includes many species in Arctostaphylos and Ceanothus, as well as the serotinous species of
Pinus and Hesperocyparis (Cupressus, see Adams et al. 2009). Generally, they take several
years to nearly 10 years before they begin producing seeds. Following that time period, seeds
that are not found by rodents and other seed predators are slowly added to the soil, eventually
reaching a depth at which they can survive a fire. Animal caching may be important to burying
seeds of manzanitas (e.g., Tevis 1953, Vander Wall et al. 2005). It is not clear how long it takes
to produce a persistent soil seed bank in which seeds are buried deep enough to survive fire. The
time between germination of obligate seeders and the production of a viable persistent soil seed
bank is considered a period of ‘immaturity risk’ for obligate seeders (Zedler 1995).
Seeds that fall to the ground suffer from high rates of seed predation (Keeley and Hays 1976,
Kelly and Parker 1990, O’Neil and Parker 2005). Models of chaparral dynamics that vary the
timing of seed survival following fire in the soil seed banks of obligate seeders indicate extreme
shifts in the dynamics of the vegetation through time (Syphard et al. 2006). Examples of native
obligate seeders in the Presidio are Arctostaphylos montana subsp. ravenii and Ceanothus
thyrsiflorus. To the north, the maritime area of Marin include Arctostaphylos virgata, A.
sensitiva on Bolinas Ridge and Mt. Tamalpais, A. montana and A. canescens on Mt. Tamalpais,
Ceanothus gloriosus varieties, Ceanothus thyrsiflorus, Ceanothus jepsonii and Ceanothus
cuneatus. To the south of the Presidio, on San Bruno Mt and in the northern part of the Santa
Cruz Mts are A. crustacea varieties, A. imbricata, A. montaraensis, A. uva-ursi, A. regismontana,
C. thyrsiflorus and C. cuneatus. Pinus muricata and other closed cone pines and cypresses in
California native within theses ranges are also obligate seeders, being killed by fire and relying
on a persistent seed bank held in serotinous cones on the trees.
Persistent Soil Seed Banks:
No seedlings have ever been seen associated with the Presidio manzanita (Recovery Plan 2003).
This results from Arctostaphylos producing seed with dormancy mechanisms that prevent their
germination until a fire has broken the dormancy (Parker and Kelly 1989). Three woody
chaparral genera disperse seed locally and create dormant, persistent soil seed banks (Keeley
1977, 1986, 1991, Parker and Kelly 1989, Kelly and Parker 1990, Zammit and Zedler 1988,
1993, 1994, O’Neil and Parker 2005). These seed are stimulated by the effects of fire, although
the mechanisms differ. For Ceanothus, seed are stimulated by a heat pulse that travels through
the soil column after a fire passes. Adenostoma seems to have some seeds that lack any
dormancy, some that respond to a heat pulse, and some that require chemicals from smoke
(Parker and Kelly 1989). A heat pulse changes an area on the seed (where it originally was
attached inside the ovary), causing a collapse of some cells and permitting water to enter the
seeds. This releases the seeds from dormancy for once water enters the seed, the seed expands
rapidly, cracking open the seed coat and permitting the embryo to germinate into a seedling. For
Arctostaphylos, the seeds do absorb a small amount of water each growing season, but do not
break dormancy and germinate unless they receive a physiological stimulus from chemicals
found in smoke.
Persistent soil seed banks are thought currently to develop slowly in Arctostaphylos. After
seedlings establish following fire, survivors generally take several years to nearly 10 years before
they begin producing seeds. Primary dispersal (gravity in Arctostaphylos as well as some coyote
and fox dispersal) adds seeds to the soil surface. Over time, seeds that are not removed by
rodents and other seed predators are slowly added to the soil seed bank, eventually reaching a
depth at which they can survive a fire. It is not clear how long it takes to produce a persistent
soil seed bank in which seeds are buried deep enough to survive fire. Seeds that fall to the
ground suffer from high rates of seed predation (Keeley and Hays 1976, Kelly and Parker 1990,
O’Neil and Parker 2005). The period of time following fire until the seeds reach a depth at which
they can survive fire has been termed “immaturity risk” (Zedler 1995) and probably exceeds 20
The model for the development of soil seed banks just described does not consider secondary
dispersal. Secondary dispersal is movement of seeds following initial dispersal, usually by
mammals in a genus like Arctostaphylos. Post-fire clusters of manzanita seedlings arising from
animal caches have been observed before in Arctostaphylos (Tevis 1953, Keeley and Zedler
1978). In the former instance (Tevis 1953) caches were thought of as a problem for re-
establishing conifers, and in the latter (Keeley and Zedler 1978), thought to be more likely due to
fusion of seeds together in the species studied (A. glauca). For Arctostaphylos species, animal
caches inadvertently act as a fire adaptation, burying seeds to depths below the heat of intense
fires (Parker, unpublished data). Secondary dispersal of seeds by animals is known to be a
critical process for a number of plants (Wenny 2001, Vander Wall and Longland 2004). Plant
population size and dynamics as well as their ability to colonize new areas may depend on the
caching of seeds by seed predators acting as secondary dispersers (Hulme and Benkman 2002).
While secondary dispersal is poorly known for Arctostaphylos, the fruits produced are dry
drupes, appropriate for animal dispersal. The mesocarp and endocarp are dry and mealy (or
lacking), respectively, and the seeds are contained inside hard, stony endocarps. Seed predation
is prevalent by rodents (Keeley and Hays 1976, Kelly and Parker 1990, O’Neil and Parker 2005)
and some studies report caching of Arctostaphylos seeds by seed predators (Smith 1942, Trevis
1953, Keeley and Zedler 1978, Smilgel and Rosenzweig 1974, Vander Wall et al. 2005a).
Establishment of Individuals:
For Arctostaphylos montana subsp. ravenii, as for all species with persistent, fire-stimulated soil
seed banks, new individuals establish only after fire. Seedlings are generally not very shade
tolerant, and are favored by the high light, high water stress environments in post-fire chaparral
compared to other species found in chaparral. Obligate sprouters (e.g, Quercus, Rhamnus or
Heteromeles) establish new individuals in closed, old growth vegetation.
Dynamics of Populations:
Obligate seeders like Arctostaphylos montana subsp. ravenii experience non-overlapping
generations at a site in the dynamics of their populations. Although there is only a single
individual remaining, in practice, ‘populations’ of the Presidio manzanita would respond like
other obligate seeders.
Fire kills off the adult population, stimulating the germination of a new cohort of individuals that
reestablish the next generation. This means that all individuals in a population are of the same
age. Any loss of individuals permanently reduces the population until the next fire. For these
types of species, the smaller the population, the smaller its range or density, the poorer its
response following the next fire.
Facultative seeders also have a persistent seed bank, but generally lose very few adult individuals
to wildfire unless the intensity is of high magnitude. Seedling response from the soil seed bank
is usually much lower than obligate seeders (Keeley and Zedler 1978) and establishment of
seedlings does not guarantee recruitment of new individuals into the population, as resprouting
individuals may remove resources too rapidly for new seedlings to successfully establish.
Overall, facultative seeders like Arctostaphylos glandulosa or A. crustacea experience only a
small percentage flux in their population from fire to fire in open chaparral habitat. The
exception comes in maritime areas in which chaparral is often slowly invaded by trees which
shade out and kill the facultative seeders. This may permanently restrict their populations or
ranges, as their response from seed banks is often poor.
Obligate sprouters (Rhamnus, Heteromeles, etc) experience a different population dynamic. In
most circumstances, only a small percentage of adults are lost in fire, and these individuals
resprout rapidly. Similar to other chaparral species, invasion by trees changes open chaparral
habitat, with loss of obligate sprouting individuals as trees invade. One difference is that they
have greater shade tolerance than obligate seeders or facultative sprouters and often persist if the
trees are not in high density, even as forest understory. Also, the shaded forest may facilitate
recruitment of new individuals in older stands of trees. Consequently, many of these species are
considered members of multiple types of plant communities. Finally, because of their shade
tolerance, they can establish beneath individuals of chaparral, such as the Presidio manzanita.
These would be the most likely types of species to provide problems in stands of Presidio
manzanita, both the original site and other locations containing cloned individuals.
V. Range, Distribution and Status of Populations
A. Historical distribution:
While there is currently only a single original individual of the Presidio manzanita remaining
overlooking the ocean above Lincoln Avenue, others have interpreted a number of collections
from San Francisco as being that of A. montana subsp. ravenii. The following table (Table 2) is
reproduced from the original Presidio Manzanita Recovery Plan (USFWS 1984). If these
specimens have been interpreted correctly, then the Presidio manzanita may have been much
more widespread throughout San Francisco on both serpentine and greenstone soils, located not
only at the Laurel Hill Cemetery (now Laurel Heights), but also at the Masonic Cemetery (now
USF, south of Laurel Heights), Mt. Davidson, and possibly near the former Protestant Orphan
Asylum (near Haight and Laguna Sts). Many of these represent disputed interpretations, for
example the specimen from Mt. Tamalpais probably represents the normal Mt. Tamalpais
manzanita (A. montana). Genetic work would be required on any specimens to distinguish it
from the Franciscan (A. franciscana) and Mt. Tamalpais (A. montana) manzanitas.
Table 2. Collections that may represent the Presidio manzanita (table modified from the Presidio Manzanita
Recovery Plan; USFWS 1984)
Locality Specimen With Franciscan Collector Year
Orphan Asylum None ? Behr 1892
Mt. Tamalpais CAS ? Eastwood 1903
Laurel Hill Cemetery UC + Brandegee 1907
Laurel Hill Cemetery UC + Brandegee 1908
Mt. Davidson CAS + Stewart 1918
Mt. Davidson JEPS + Blasdale 1923
Masonic Cemetery CAS + Clark 1928
Laurel Hill Cemetery UC + Edmunds 1938
Laurel Hill Cemetery UC + Wieslander 1938
While some of these interpretations may be questionable, the overall impression is that the
distribution of the Presidio manzanita was probably much more widespread throughout San
Francisco on serpentines and greenstones before development and the planting of exotic trees
eliminated these habitats. From collections examined, the Presidio manzanita may well have
occurred outside the Presidio, at least to Mt. Davidson. One of the Mt. Davidson collections
(Stewart 1918 CAS 38903) exhibits leaves that are much rounder than other specimens, similar
to that found on the Presidio manzanita. Other specimens, such as one from the Masonic
Cemetery (Clark 1928 CAS 159815) shows leaves with a length to width ratio closer to A.
montana subsp. ravenii than to other collections of A. franciscana. It is not possible to
determine ploidy levels from these herbarium collections with current technology, but it may be
possible to extract DNA and sequence for a few genes for which the Presidio and Franciscan
manzanita differ; that potential depends on the integrity of the DNA remaining in the specimens.
B. Current distribution of individuals of the Presidio Manzanita:
Near the WWII monument on Lincoln Blvd in the Presidio is the remaining individual of the
Presidio manzanita. Associated with that individual are a number of clones that were planted
adjacent to the mother plant, and have successfully established (Figure 9). Eight of these
individuals were planted by GGNRA staff in 1987, with two subsequently planted in 1988. Nine
are still present as of the last survey (2008) and the plants have not changed status for the past 5
years (see Appendix 1 as well).
Figure 9. Population of A. montana subsp. ravenii at original site. Mother plant with clones. In this figure, the
mother plant and one of the clones are labeled. Photo from summer 2009. Photo by V.T. Parker.
Near this site, several clones were also planted in 1987 in a site called the Triangle (bounded by
Lincoln and Kobbe Ave) and one individual (that appears like two individuals) remain. Also
nearby, in a site called the Yerba Buena serpentine site, seven individuals have established and
appear to be expanding (Fig 10-11). Twenty-two clones were originally planted at the Yerba
Buena site in 1987 by GGNRA staff; by 1994, 12 were remaining. This rate of survival at all
these sites should be considered excellent because the 6 years following the planting of these
clones were severe drought years. The climate fluctuated thereafter, with wet years interspersed
with drought years. By 1998, approximately 7 or 8 of the clones at the Yerba Buena site had
stabilized and remain today.
Figure 10. Yerba Buena serpentine site with a population of Presidio manzanita clones. Photo by V.T. Parker.
Most of the individuals at the Yerba Buena site are healthy and expanding. Seven individuals
are spread about in a mixed maritime chaparral, coastal scrub, and coastal prairie site. An
example individual is shown in Figure 11.
Figure 11. Example of one of the Presidio manzanita clones in the Yerba Buena site. Photo by V.T. Parker.
Other locations of clones of the Presidio manzanita:
A summary of all known locations of clones of the Presidio manzanita are found below (Table
3). Plantings had been made at Inspiration Point, West Crissy Bluffs and along the path of the
WWII memorial. Failure of their survival relates to a number of local unique issues, or are
unknown, for example, off-road bicycles (IP), unknown (West Crissy Bluffs), and horticultural
adjustment of the site (removal and replacement of any surviving individuals following some
mortality) (WWII). Surviving individuals are at botanical gardens or nursery locations.
Table 3. Locations of all known clones of the Presidio manzanita and their current status.
C. Threats to Populations:
Most of the threats to these populations were discussed well in the 2003 Recovery plan. These
threats continue to some extent. The principal threats to bring up again are 1) the need to have
back-up populations. Castastrophic climatic variation, earthquakes, fire or other catastrophes are
all possible. Having a number of back up populations in other locations is fairly critical. Having
some away from the San Francisco Bay Area is also critical. This could be having a number of
clones of the mother plant and of seedlings planted at another botanic garden such as Santa
Barbara Botanic Garden or Rancho Santa Ana Botanic Garden, both in southern California.
Another continuing issue is invasive plants and encroaching native plants. The area around the
mother plant appears to be well tended with respect to introduced species that may risk the
Presidio manzanita. However, a number of native species are also present that are large enough
to create problems for these prostrate plants. As an example, soap-plant (Chlorogalum
pomeridianum) individuals are quite large in association with the Presidio manzanita and where
they have bloomed and expanded leaves, large diebacks are coincident in the manzanitas (Figure
12). This pattern also occurs with other species, such as gumweed (Grindelia hirsutula subsp.
maritima) to a lesser extent.
Figure 12. Example of diebacks caused by shading from other species, June 2009. Photo by V.T. Parker.
This problem is likely to be more difficult to manage as populations are spread about because of
different combinations of species. At the Yerba Buena site, for example, the plant assemblage
exhibits elements of maritime chaparral, coastal scrub and coastal prairie. Some invasive species
are present at the site, such as introduced Cotoneaster and Rubus species. An example of
problems likely to arise with associated species is illustrated below (Figure 13).
An additional threat includes fungal pathogens that may cause diebacks during exceptionally wet
winters (see 2003 Recovery Plan, p. 69). In the past there have been diebacks attributed to
fungal infections. The dieback observed previously was attributed to species of Phomopsis that
affects manzanitas (2003 Recovery Plan, p. 70). Normally, manzanitas can recovery from
temporary diebacks as long as conditions change back in sufficient time. The pathogen involved
in sudden oak death, a species of Phytopthora appears to be common among Ericaceous species,
but not yet observed causing dieback (2003 Recovery Plan, p. 70).