ArticlePDF Available

MALACOTHAMNUS ENIGMATICUS (MALVACEAE), A NEW RARE SPECIES FROM THE DESERT EDGE OF THE PENINSULAR RANGE IN SAN DIEGO COUNTY, CA

Authors:

Abstract and Figures

Malacothamnus enigmaticus K.Morse & T.Chester (Malvaceae) is a new species endemic to the desert edge of the Peninsular Range in San Diego County, California, USA, from the San Ysidro Mountains in the north to the Laguna Mountains in the south. It is the only Malacothamnus species in its geographic range, which contains a unique habitat in San Diego County. Malacothamnus enigmaticus is morphologically distinct in many inflorescence characteristics from its two most similar species, M. densiflorus (S.Watson) Greene and M. fasciculatus (Nutt. ex. Torr. & A.Gray) Greene, which are also geographically closest to the new species. A Principal Components Analysis (PCA) of morphological features shows that M. enigmaticus is clearly distinct from both M. densiflorus and M. fasciculatus, and that all three species show the same level of separation from each other. Hence M. enigmaticus should be equally recognized at species rank. Malacothamnus enigmaticus and M. fasciculatus are both distinguished from M. densiflorus by their much-higher density of stellate stem hairs, and in most specimens by their shorter stellate hair rays and more numerous stellate hairs on the calyx. Malacothamnus enigmaticus is distinguished from M. fasciculatus by its longer calyx bracts, and generally by its wider stipular bracts. Malacothamnus enigmaticus is a rare species, known from just two ~150 km2 areas separated by a gap of ~15 km, one in the Laguna Mountains and one in the Pinyon Ridge / Culp Valley / San Ysidro Mountain area. Due to its limited range and small populations, we propose a California Rare Plant Rank of 1B.3 for the species.
Content may be subject to copyright.
MADRO ˜
NO, Vol. 66, No. 3, pp. 103–119, 2019
MALACOTHAMNUS ENIGMATICUS (MALVACEAE), A NEW RARE SPECIES FROM
THE DESERT EDGE OF THE PENINSULAR RANGE IN SAN DIEGO COUNTY, CA
KEIR MORSE
Rancho Santa Ana Botanic Garden, 1500 North College Ave., Claremont, CA 91711
kmorse@rsabg.org
TOM CHESTER
1802 Acacia Lane, Fallbrook, CA 92028
ABSTRACT
Malacothamnus enigmaticus K.Morse & T.Chester (Malvaceae) is a new species endemic to the desert edge
of the Peninsular Range in San Diego County, California, USA, from the San Ysidro Mountains in the north
to the Laguna Mountains in the south. It is the only Malacothamnus species in its geographic range, which
contains a unique habitat in San Diego County. Malacothamnus enigmaticus is morphologically distinct in
many inflorescence characteristics from its two most similar species, M. densiflorus (S.Watson) Greene and M.
fasciculatus (Nutt. ex. Torr. & A.Gray) Greene, which are also geographically closest to the new species. A
Principal Components Analysis (PCA) of morphological features shows that M. enigmaticus is clearly distinct
from both M. densiflorus and M. fasciculatus, and that all three species show the same level of separation from
each other. Hence M. enigmaticus should be equally recognized at species rank. Malacothamnus enigmaticus
and M. fasciculatus are both distinguished from M. densiflorus by their much-higher density of stellate stem
hairs, and in most specimens by their shorter stellate hair rays and more numerous stellate hairs on the calyx.
Malacothamnus enigmaticus is distinguished from M. fasciculatus by its longer calyx bracts, and generally by
its wider stipular bracts. Malacothamnus enigmaticus is a rare species, known from just two ~150 km
2
areas
separated by a gap of ~15 km, one in the Laguna Mountains and one in the Pinyon Ridge / Culp Valley / San
Ysidro Mountain area. Due to its limited range and small populations, we propose a California Rare Plant
Rank of 1B.3 for the species.
Key Words: bracts, Malacothamnus aboriginum,Malacothamnus densiflorus,Malacothamnus fasciculatus,
morphology, taxonomy.
Malacothamnus Greene is a genus of shrubs and
subshrubs endemic to the California Floristic Prov-
ince, the adjacent desert transition zone, and a
disjunct area of central Arizona having many taxa
in common with the California Floristic Province.
The genus is well-known as a fire-follower with
plants germinating abundantly post-fire, flowering as
early as the first year post-fire, and often dying off as
soon as three to four years post-fire due to
competition with other shrubs (Bates 1963). Mala-
cothamnus taxa are notoriously difficult to separate,
with ‘‘a lack of distinct morphological characters to
delineate taxa’’ complicated by ‘unclear . . . taxo-
nomic boundaries among the species . . . since
morphology within species can be highly variable’’
(Slotta 2004). Furthermore, molecular analysis has so
far been unable to distinguish between species (Slotta
2004).
All recent floral treatments (Munz 1974; Beau-
champ 1986; Bates 1993; Slotta 2004, 2012; Bates
2015) agree that at least two species of Malacotham-
nus grow in San Diego County, California: M.
fasciculatus (Nuttall ex Torrey & A.Gray) Greene
and M. densiflorus (S.Watson) Greene. Bates (1963)
was first to recognize a third species from the Laguna
Mountains in San Diego County, which he consid-
ered indistinguishable from M. aboriginum (B.L.Ro-
binson) Greene except for two characteristics of its
calyx bracts (called involucral bracts by Bates). Note
that the San Diego County Laguna Mountains
should not be confused with Laguna Mountain in
San Benito County where M. aboriginum actually is
found. All references to Laguna in this paper refer to
the range in San Diego County. A gap of 540 km lies
between the San Diego County populations and the
closest known M. aboriginum specimen in Kings
County.
All treatments since Bates (1963) have called the
Laguna plants M. aboriginum, except for Slotta
(2004, 2012). She removed the Laguna plants from
the range of M. aboriginum but was silent on where
they should be placed. The most recent treatment
(Bates 2015) says plants ‘‘in the Laguna Mountains
in San Diego County... appear to be M. aboriginum’’
(emphasis added) and that they sometimes have been
assigned to M. densiflorus.
Only two morphometric analyses of Malaco-
thamnus have been done (Benesh and Elisens 1999;
Slotta 2004), and neither assessed the Laguna
plants. We decided to conduct a similar analysis
that included the Laguna plants to resolve their
species placement.
Examining specimens of all Malacothamnus taxa,
we found that the Laguna plants were unique in
their combination of inflorescence type, bract
characteristics, and pubescence, with morphology
most like those closest in range to it: M. fasciculatus
and M. densiflorus. Focusing on these species for
comparison, we measured a wide variety of speci-
mens from the full geographic range of M. densi-
florus; the full range of what was circumscribed as
M. fasciculatus var. fasciculatus by Slotta (2004); all
specimens suspected to be similar to the new taxon
from the Lagunas, the immediately-surrounding
areas, and throughout Baja California (many
annotated as M. aboriginum); as well as two
specimens of M. aboriginum from San Benito
County. See Fig. 1 for the geographic locations of
the measured specimens in the San Diego County
area. This paper describes the Laguna plants as a
new species, M. enigmaticus K.Morse & T.Chester,
andusesthisnamefromhereforwardformaximum
clarity.
MATERIALS AND METHODS
Focusing on the reduced set of taxa as noted
above, we measured the characteristics found most
useful in the previous two morphometric studies,
concentrating on the calyx bracts, calyx, stem and
leaf pubescence, and inflorescence structure. Ad-
ditional characteristics not used in previous studies
included stipular bracts and glandular hairs. We
refined stem hair density from the qualitative
measures used in previous studies to a quantitative
measure after Principal Components Analysis
(PCA) comparison of preliminary data indicated
much greater utility of the quantitative measure
for taxon delineation. Some measurements re-
mained categorical, such as the number of flowers
per stem node which could not be accurately
counted without significant damage to many
specimens since they are often tightly clustered
and overlapping (Appendix 1). Our final set of
measurements included 50 characteristics (Appen-
dix 2). Because bract terminology in previous
treatments has been inconsistent, vague, and/or
confusing, we explicitly define our terms below.
We also separately discuss how the stem pubes-
cence was measured.
We measured 152 specimens, of which 71 were
gathered from the field (see details below) and 81
were herbarium specimens from SD, UCR, and
RSA (see Fig. 1 for the geographic locations of the
measured specimens in the San Diego County area).
Measured specimens were from Los Angeles,
Orange, Riverside, San Benito, and San Diego
counties in California, as well as from Baja
California, Mexico. Sampled specimens were chosen
to represent the maximum morphological variation
of each of these taxa, and to give good geographic
sampling of each taxon. Specimens without mature
measurable calyxes and bracts were avoided as were
a few with significant damage to the stem hairs.
Some collections from the same plant and/or
population were measured to show variation within
each.
Field Work
We surveyed for Malacothamnus on 52 separate
days from 13 February 2015 to 14 December 2017 to
find and study plants in the field. We made 128
specimen collections on 28 separate days. In many
cases, we took three or more separate sheets at a
given location to represent the maximum morpho-
logical variation at a site, usually the least and most
robust inflorescences as well as a typical one,
sometimes all from the same plant. Several field trips
were devoted to delineating the distribution of M.
enigmaticus, with four trips specifically made to
record its characteristics in the field, where we
assessed the habitat and measured plant size, shape/
form, and lower leaves, characteristics that cannot be
assessed from herbarium specimens.
Details on the Measured Characteristics – Bract
Terminology
There has been general agreement that there are
leaves (sometimes called leaf-like bracts) subtended
by stipules in the Malacothamnus inflorescence, and
that the calyx is subtended by a whorl of three bracts
that have been given various names by different
authors. However, the intervening bracts are not
clearly defined in any of the previous treatments. In
this section, we define the terminology we use in this
paper. Our terminology is simplified from previous
authors in the hopes of reducing confusion. See Fig.
2 for a labeled diagram illustrating the terminology
we are using.
Directly below the calyx are three bracts forming a
whorl (Fig. 2A and 2B). These have collectively been
called the involucel or epicalyx. Individually they
have been called bracts, bractlets, bracteoles, involu-
cral bracts, involucellar bractlets, and involucellar
bracteoles. We call them an involucel made up of
three calyx bracts.
Leaves in Malacothamnus generally decrease in
size distally and fairly large single leaves can often be
found in the inflorescence, one at each node.
Technically, leaves in the inflorescence are bracts,
but there is often no clear line in Malacothamnus
where leaves end and leaf-like bracts start. A new
inflorescence branch may begin at any of the
uppermost leaves. This often makes the boundary
between inflorescence branches and axillary inflores-
cences unclear. Since there already are an abundance
of bract types in the inflorescence, for clarity we call
them leaves if they look like leaves, have stipules, and
are unmodified from the lower leaves beyond
reduction in size. Malacothamnus palmeri (S.Watson)
Greene is the only species in the genus to have leaves
in the inflorescence that are modified in shape as well
as size; we call those ‘‘leafy bracts’’.
Subtending each leaf is a pair of stipules which are
often leafy, but sometimes reduced to scales.
Occasionally these stipules are basally fused in the
distal end of the inflorescence. Vestigial leaves may
be found with the more-proximal fused stipules, but
104 [Vol. 66
MADRO ˜
NO
often only persist for a few nodes. Hill (1982) noted
this in the related genus Malvastrum:
‘‘In many individuals a progression can be clearly
seen from well-developed leaves with associated
stipules, to reduced leaves with stipules, stipules
with a leaf remnant attached, and finally to a bifid
bract.’’
This is illustrated in Fig. 2D. When a bifid bract is
present at the base of a node on the main
inflorescence axis in the stipular position, we call
these ‘‘fused stipules’’.
At a given node, additional bracts resembling
stipules or fused stipules are often present between
the stipules and the calyx bracts. These bracts have
not been clearly defined in any Malacothamnus
treatment. In Malvaceae, similar bracts have been
called floral bracts, stipular bracts, and/or inflores-
cence bracts (S. Hill, University of Illinois, personal
communication). Here we use the term ‘‘stipular
bracts’’. In M. enigmaticus, clusters of stipular bracts
subtend clusters of flowers in the inflorescence. The
shape of these stipular bracts can vary considerably
even within a single glomerule of the inflorescence
(Fig. 2C and 2E).
Details on the Measured Characteristics – Hairs
Malacothamnus hairs often change along the stem,
especially near the distal end of the inflorescence
where they may be sparser, longer-rayed, and/or
more stipitate. For consistency and to avoid disparity
in developmental maturity at the distal end of the
inflorescence, we measured the stem hairs near the
proximal end of what is generally found on
specimens. We measured stem hair type (stellate or
simple glandular), density, length of stellate hair rays,
length of stellate hair stipes, and length of glandular
hairs.
Stellate stem hair density was sampled in a 2.0 3
2.0 mm square where the stem was 2–4 mm (generally
3 mm) in diameter. The square we actually used for
measurement was 2.0 32.3 mm. To avoid confusion,
we refer to it as a 2 32 mm square in this paper, and
all numbers given here were scaled from the actual
measurements by multiplying by 2.0 / 2.3.
Condition and size of the specimen, as well as ease
in accessing a sampling location, often influenced
where we measured the stem. Areas where hairs were
damaged, or not typical of that region of the stem,
were avoided. A card with a 2 32 mm square hole
FIG. 1. The full geographic range of M. enigmaticus (filled squares), along with nearly the full geographic range of M.
densiflorus (open circles), and the nearby portion of the range of M. fasciculatus (asterisks). Measured and analyzed
specimens are shown with larger symbols. Malacothamnus enigmaticus likely exists at or near the single location marked with
a question mark (see text). The Malacothamnus specimens in northeast Baja California (open stars) appear to differ from
each of these three species, but too few specimens are known to generate reliable conclusions.
2019] 105
MORSE AND CHESTER: MALACOTHAMNUS ENIGMATICUS, A NEW RARE SPECIES
was placed over the sample area, the hairs were
gently scraped off within this square, and then
counted on another surface. In some species, such
as M. fasciculatus and M. enigmaticus, the hairs may
be very small and in densely-overlapping layers. In
these cases, care was taken to try and remove as
many individual hairs as possible (rather than
removing clumps). When scraped off in clumps,
FIG.2. Malacothamnus enigmaticus bracts and a dissected glomerule. A. Basal view of flower showing calyx and the three
subtending linear calyx bracts. B. Side view of calyx and calyx bracts. C. Some variation in stipular bracts. D. Transition
from stipules with reduced leaf (bottom/proximal) to fused stipules (top/distal). E. Example of a dissected glomerule with a
reduced leaf (1) and stipules (2) at the base transitioning to two reduced branches with alternating clusters of stipular bracts
(3) subtending clusters of flowers (4). The calyx bracts are shown attached to the calyx. Illustrations by Elysa DuCharme.
106 [Vol. 66
MADRO ˜
NO
hairs can be much more difficult to separate and
count.
Stem stellate hair rays were measured from the 2 3
2 mm sample and surrounding areas. Stem stellate
hair stipes were measured from the areas on the stem
immediately surrounding the 2 32 mm sample area.
When scraped, the stipes often broke above the base,
so stipes were only measured for hairs still attached
to the stem. Stem glandular hair length was measured
near the 2 32 mm sample area, but with more
flexibility in distance in order to record a greater
range of hair lengths, as they were much sparser than
the stellate hairs.
The stem hairs from a representative specimen of
each of the four species analyzed were also examined
with a scanning electron microscope (SEM). Stem
sections of the selected species were mounted on
aluminum stubs using conductive tape, coated with
gold using a Cressington Sputter Coater 108auto
system (Cressington Scientific Instruments, Watford,
UK), and imaged on a Hitachi SU3500S Variable
Pressure SEM (Hitachi High Technologies America,
Inc., Pleasanton, CA) at Rancho Santa Ana Botanic
Garden.
Principal Components Analysis
Our fundamental analytical tool was a principal
components analysis (PCA) performed using the
princomp function in R (R Core Team, R Founda-
tion for Statistical Computing, Vienna, Austria),
which removes the mean from each measurement,
and then divides the result by its standard deviation.
A PCA is a powerful tool to analyze a dataset
composed of numerous measurements of a large
number of specimens to easily visualize whether the
specimens group into separate taxa or not; see, e.g.,
Keeley et al 2007. A PCA linearly transforms a
dataset with multiple, usually highly correlated,
variables into a new set of uncorrelated variables
ranked in order of how much variance in the dataset
is encompassed by each variable in turn. Typically,
most of the variance in the data is explained by the
first few ‘‘principal components’’, allowing easy
visualization of the structure of the original data set
using just a few variables. Another advantage of a
PCA is that it is an unbiased approach, weighting all
measured characteristics equally, to studying whether
multiple species are supported by the data, in
contrast to selecting ‘‘key’’ characteristics which
may or may not be supported by differences in other
characteristics.
We began our analysis after measuring just 19
characteristics on the first 15 specimens. We used the
preliminary analyses to guide us in selecting addi-
tional characteristics to measure and in selecting
additional specimens to measure. We strove to
sample all the variation we saw in the voucher
specimens we examined and, as detailed above, in the
plants in the field. It proved very useful to have
multiple measured specimens from the same plant to
provide a quantifiable baseline of variation within a
single species. This drove us to measure additional
characteristics, and/or to sharpen our measurements
of some characteristics, to try to separate the species
more cleanly. We also found that the inflorescence
length and the lengths of the internodes above the
most proximal internode were not useful in separat-
ing the species and eliminated them in the PCA.
Those lengths are heavily influenced by the timing of
the collections since the distal internodes elongate
much later than proximal internodes. We retained
the length of the most proximal internode since it
matures first and is fairly consistent in each species.
We initially analyzed all specimens of all species
together, but soon found that it was better to analyze
just two or three species at a time, especially those in
a given geographic area, since a plot of the first two
PCA components readily discriminates two or three
species, but often produces significant overlaps with
more species.
RESULTS
Visual comparison of specimens of M. enigmaticus
to the rest of genus found M. enigmaticus to be most
similar in morphology to M. densiflorus and M.
fasciculatus, rather than to M. aboriginum as
suggested by Bates. Bates (1963) considered M.
aboriginum indistinguishable from M. enigmaticus
with the exception of the calyx bracts being ‘‘mostly
2–9 mm broad and often connate basally’’ in M.
aboriginum compared to ‘‘rarely more than one
millimeter broad and free’’ in M. enigmaticus. While
both M. aboriginum and M. enigmaticus are united in
the uncommon characteristic of wide stipular bracts,
which may have caused Bates to place them together,
they are widely divergent in two other important
characteristics: pubescence type and calyx lobe
shape. Malacothamnus aboriginum has large and
sparse stellate hairs with very wide, subcordate calyx
lobes, whereas M. enigmaticus has minute and dense
stellate hairs with much narrower, ovate calyx lobes.
Scanning electron micrographs of representative
stem hairs of the three San Diego County Mala-
cothamnus species and M. aboriginum from San
Benito County are shown in Fig. 3. Malacothamnus
enigmaticus and M. fasciculatus have a relatively-high
density of hairs (Fig. 3A and 3B). In and near San
Diego County, the stellate hair rays are often longer
in M. fasciculatus than M. enigmaticus, but more
northern and higher elevation M. fasciculatus are
comparable to M. enigmaticus in ray length. Mala-
cothamnus densiflorus and M. aboriginum have a
relatively low density of hairs (Fig. 3C and 3D). Note
that the ray length of M. aboriginum vastly exceeds
those of the other taxa (Fig. 3D).
As expected, the PCA with all species together
showed significant separation of M. aboriginum from
the remaining species (Fig. 4). Hence M. aboriginum
was excluded from all additional PCAs. In this PCA
with all species together, PCA1 accounts for 32% of
2019] 107
MORSE AND CHESTER: MALACOTHAMNUS ENIGMATICUS, A NEW RARE SPECIES
the total variance; PCA2 for 13%; and PCA3 for
9%. PCAs not shown indicated some specimens in
Baja California are possibly undescribed taxa which
need further sampling to adequately assess. As a
result, these Baja California specimens were not used
in the PCAs shown here.
The PCA results of the 138 remaining specimens
(Fig. 5) and the geographic map (Fig. 1) indicate that
M. enigmaticus is clearly distinct from M. densiflorus
and M. fasciculatus both morphologically and
geographically. The PCA separation is robust and
comes from many different characteristics; half of the
coefficients for the first Principal Component (PCA1)
in both analyses are within a factor of two of the
largest coefficient (Appendix 2). Those 138 specimens
span the full range of M. enigmaticus and M.
densiflorus, as well as the full range of what is
circumscribed as M. fasciculatus var. fasciculatus in
Slotta (2004), excluding what appear to be unde-
scribed taxa in Baja California.
The morphological separation of M. enigmaticus
and M. densiflorus (Fig. 5A) is especially good, with a
gap between all members of each of the species in the
PCA plot. Furthermore, the PCA shows that the set
of measured plants of M. enigmaticus is homoge-
neous, since the dispersion in the PCA values for M.
enigmaticus is essentially the same as the dispersion in
the PCA values of three samples from one plant (the
FIG. 3. Scanning electron micrographs showing a 1 31 mm square of stem hairs near the base of the inflorescence typical
of the three San Diego County Malacothamnus species and M. aboriginum from San Benito County (Photo: K. Morse). A.
M. enigmaticus (Morse M023). B. M. fasciculatus var. fasciculatus (Morse M007). C. M. densiflorus (Morse M020). D. M.
aboriginum (Morse M110).
108 [Vol. 66
MADRO ˜
NO
diamond points in the plot). These samples were
intentionally chosen to represent the extreme amount
of variation within a single plant (more than typical),
as we wanted to compare those values across the
PCA plot. A separate PCA using only M. enigma-
ticus specimens (not shown here) produced the
spherical cloud of points, with no geographic trend,
expected if the species is homogeneous.
In contrast, M. densiflorus is heterogeneous. The
samples from the same plant (Fig. 5A, the points
with separate symbols in the plot) occupy only a
small range of the variance in its PCA. A separate
PCA using only M. densiflorus specimens (not shown
here) shows clear trends with geography, which
roughly correlates with its two previously recognized
varieties and possibly a third undescribed taxon.
The presence of subtaxa within M. densiflorus
skews the PCA plot in Fig. 5A such that the two
species are separated along a diagonal in the plot. If
only two species were present, each of which is
homogeneous, they would be primarily separated
along the PCA1 axis.
The morphological separation of M. enigmaticus
and M. fasciculatus (Fig. 5B) is excellent. Again, M.
enigmaticus is seen to be homogeneous, even though
it is spread over quite a range in PCA1 values, since
the separate samples from the same plant span
almost the full range of PCA1 values for the
specimens sampled.
Malacothamnus fasciculatus is heterogeneous,
again seen from the samples from the same plant
occupying only a small range in its variance in PCA1
and PCA2 (Fig. 5B). A separate PCA using only M.
fasciculatus specimens (not shown here) shows clear
trends with geography. Although the trends do not
seem to correlate with previously defined varieties of
the species, they indicate that there is variation within
M. fasciculatus that may warrant further circum-
scription.
The presence of subtaxa within M. densiflorus and
M. fasciculatus, and the resulting different structure
of the two PCA plots in the comparison of each with
M. enigmaticus, accounts for the difference in
variance explained by the first two principal compo-
nents. PCA1 accounts for 38% of the total variance
between M. enigmaticus and M. fasciculatus, cleanly
separating the taxa with just PCA1 (Fig. 5B). In
comparison, PCA1 accounts for just 24% of the total
variance between M. densiflorus and M. enigmaticus,
since the separation is along a diagonal line relating
PCA1 and PCA2 (Fig. 5A). The total variance
accounted for by both PCA1 and PCA2 combined
is similar in the two PCAs, 47% for M. enigmaticus
and M. fasciculatus and 43% for M. densiflorus and
M. enigmaticus. This total variance is much higher
than that from a random distribution derived from
the broken-stick model (Jackson 1993); the broken-
stick model gives that the variances of the first nine
PCA components are better than expected from
FIG. 4. Principal Component Analysis using all 50 measured characteristics for 140 Malacothamnus specimens, 138 from
the San Diego County area northward to Los Angeles and San Bernardino counties, plus two specimens of M. aboriginum
from San Benito County. The unique characteristics of M. aboriginum were best captured by the third principal component,
where it is separated from the other three species by roughly 13 standard deviations.
2019] 109
MORSE AND CHESTER: MALACOTHAMNUS ENIGMATICUS, A NEW RARE SPECIES
FIG. 5. Principal Component Analysis (PCA) for 138 Malacothamnus specimens from the San Diego County area
northward to Los Angeles and San Bernardino counties. A: PCA using all 50 measured characteristics for only specimens of
M. densiflorus and M. enigmaticus. B: PCA using 44 measured characteristics for only specimens of M. enigmaticus and M.
fasciculatus (the other six measured characteristics were identical for all specimens of these two species). The points for
specimens from the same population, and sometimes the same plant, where we intentionally sampled the extreme
morphological variation with multiple specimens, are marked with additional symbols that are unique for each population,
such as the diamonds surrounding the three M. enigmaticus data points that come from the same plant.
110 [Vol. 66
MADRO ˜
NO
chance for the M. enigmaticus and M. fasciculatus
analysis, and the variances of the first 11 components
are better than expected from chance for the M.
densiflorus and M. enigmaticus analysis.
The PCA plot shows that some specimens of M.
enigmaticus are close to some specimens of M.
fasciculatus in the values of PCA1 and PCA2.
However, the taxa are cleanly separated in some
higher-order PCA components, are disjunct geo-
graphically, and they are readily distinguished using
the key presented in this paper. Hence their species
assignment is unambiguous. In fact, one of those
specimens of M. enigmaticus has two other samples
from the same plant plotting in the middle and
farthest part of the M. enigmaticus distribution
relative to M. fasciculatus (see the specimens
represented with diamonds in Fig. 5B).
The PCA coefficients from the PCA using all three
species show that the most-important separating
characteristics of the three species are stem hair
density, the length of the longest calyx bracts, and the
width of the widest stipular bracts. The longest calyx
bract in each specimen is plotted vs. the density of
stem hairs in Fig. 6, showing that the three species
separate quite well just using two measured charac-
teristics. The widest stipular bract is plotted vs. the
longest calyx bract in Fig. 7 for M. enigmaticus and
M. fasciculatus, showing the excellent separation of
those two species using just two characteristics.
The majority of the measured characteristics show
strong differences in their mean values between at
least two of the three species. The means for 39
characteristics differ between at least two species at
the five standard deviation level or higher, with 19 of
those above the ten standard deviation level. The
strongest discriminant is the stem hair density, with
the mean values for M. densiflorus and M. fascicu-
latus differing by 21 standard deviations; M. densi-
florus and M. enigmaticus differing by 35 standard
deviations; and M. fasciculatus and M. enigmaticus
differing by seven standard deviations.
Table 1 provides some of the most important
characteristics that separate M. densiflorus,M. enigma-
ticus, and M. fasciculatus, and the range of our
measurements for each species. The measurements
came from 49 specimens of M. densiflorus,27specimens
of M. enigmaticus, and 62 specimens of M. fasciculatus.
More M. densiflorus and M. fasciculatus specimens were
measured since those two species have a much larger
geographic range and more morphological variation
than does M. enigmaticus, possibly due to the presence
of subtaxa within those two species (See Figshare Data
Repository measurements on these specimens, https://
doi.org/10.6084/m9.figshare.8061692).
Several significant distinguishing features in cate-
gorical measurements were found. The density of
hairs on the sepals and leaves averaged one category
lower in M. densiflorus than in the other two species.
The presence of glandular hairs greater than or equal
to 0.1 mm on the stem only occurs in one possible
subtaxon of M. densiflorus.Malacothamnus densi-
florus sensu stricto (s. s.), which has populations
closest in range to M. enigmaticus, has sparse hairs
near the base of the calyx, whereas M. fasciculatus,
FIG. 6. Plot of the longest calyx bract in each specimen vs. the density of stem hairs. This plot, using just two
characteristics, essentially reproduces the separation of the three species from the PCA with 50 characteristics.
2019] 111
MORSE AND CHESTER: MALACOTHAMNUS ENIGMATICUS, A NEW RARE SPECIES
M. enigmaticus, and the other possible subtaxa of M.
densiflorus have dense hairs.
CONCLUSION
Malacothamnus enigmaticus is taxonomically dis-
tinct and merits designation at the rank of species. It
occurs with no other Malacothamnus taxa and consists
of a morphologically-homogeneous group of plants
with a well-defined geographic range occupying a
unique habitat in San Diego County, where the
eastern edge of montane elevations transition rapidly
to the low desert. It is morphologically distinct from
all other Malacothamnus species in its unique combi-
nation of inflorescence type, bract characteristics, and
pubescence, with morphology most like those closest
in range to it, M. fasciculatus and M. densiflorus.
The separation of M. enigmaticus from these species
is robust. It is cleanly distinguished using just two
characteristics. Furthermore, it is clearly distinguished
in a PCA even when the PCA is restricted to using
various independent subsets of the measured charac-
teristics, showing that it is distinct from those species
in a number of characteristics.
FIG. 7. Plot of the longest calyx bract vs. widest stipular bract for M. enigmaticus and M. fasciculatus. This plot, using just
two characteristics, essentially reproduces the separation of the two species from the PCA with 50 characteristics.
TABLE 1. IMPORTANT CHARACTERISTICS SEPARATING M. DENSIFLORUS,M. ENIGMATICUS,AND M. FASCICULATUS.
Values in bold type indicate characteristics significantly different (exceeding two standard deviations) from those of the
other two species.
1
The median value varies in the different possible subtaxa
2
Longest-rayed stellate hairs are generally
sparsely scattered on the stem in these three species.
Characteristic
M. densiflorus M. enigmaticus M. fasciculatus
min median max min median max min median max
Stem hair density # per 2 32 mm square 26 48 75 164 187 235 96 143 283
Stem stellate hair stipe length (mm) 0 0.1 0.4 0000 0 0.2
Stem stellate hair ray length (mm) 0.1 0.2–0.3
1
4.1
2
0.1 0.1 0.5
2
0.1 0.1–0.2
1
1
2
Most proximal infl. internode length (mm) 12 58 130 20 51 75 83075
Stipular bract width (mm) 0.5 0.7 3.0 0.5 3.0 8.0 0.5 0.7 4.0
Calyx bract length (mm) 5 8 20 3813138
Calyx bract hairs longest average (mm) 0.5 1.5 3.0 0.5 1.0 1.5 0.5 0.5 1.0
Calyx length (mm) 6.0 10.3 21.0 6.0 10.0 15.0 5.0 7.4 10.0
Calyx lobe length (mm) 3.0 6.3 13.5 3.0 6.3 10.0 1.5 3.7 7.0
Calyx lobe width (mm) 2 3.5 5.5 2 3.7 5 2 3 4
Longest calyx hair rays (mm) 1.0 2.0 4.0 0.5 1.0 2.0 0.3 0.4 1.0
112 [Vol. 66
MADRO ˜
NO
TAXONOMIC KEY TO THE SPECIES OF
MALACOTHAMNUS IN SAN DIEGO,ORANGE,AND
RIVERSIDE COUNTIES,CA
The following key unambiguously separates all the
specimens measured for this paper despite the small
overlap in each of the two characteristics used to
separate M. fasciculatus and M. enigmaticus (Figs. 6
and 7). Stem hairs are measured near the base of the
inflorescence where the stem is 2–4 mm wide. Stipular
bracts are most easily seen proximally in glomerules,
especially when flowers are immature (Fig. 8D).
1. Stem hairs 26–75 in a 2 mm x 2 mm square,
green stem generally visible to the naked eye
between hairs . ..............M. densiflorus
10Stem hairs 164–235 in a 2 mm x 2 mm square,
stem yellow or white due to densely overlapping
hairs, green stem rarely visible distally . . . . . . 2
2. Longest calyx bracts 2.5–6.0 mm; widest stipular
bracts 0.5–2(4) mm wide . . . . . . . M. fasciculatus
20Longest calyx bracts 5.5–13 mm; widest stipular
bracts 1–8 mm wide ..........M. enigmaticus
We did not include Baja California in the key since
M. densiflorus and M. fasciculatus s.s. do not extend
very far south of the San Diego County border. Our
analyses indicate much of what has been called M.
densiflorus and M. fasciculatus in Baja California
may differ from those species and requires further
investigation.
To the north of the region of this key, Los
Angeles, San Bernardino, Santa Barbara, and
Ventura counties all have M. fasciculatus. A sam-
pling of additional specimens of M. fasciculatus from
this area all fit within the key. These counties were
left out of the key due to currently-unresolved
morphological differences between M. fremontii
(Torr. ex A.Gray) Greene and M. orbiculatus
(Greene) Greene.
TAXONOMIC TREATMENT
Malacothamnus enigmaticus K.Morse & T.Chester,
sp. nov.—Type: USA, California, San Diego
County, Laguna Mountains, southernmost hill in
Garnet Peak vicinity, in sandy soil of disturbed
chaparral (fuel break); 32.9098N, 116.4638W,
1650 m, 23 June 1977, Fred T. Sproul 170
(holotype: SD114895).
Bract terminology used here is explicitly defined in
the Materials and Methods section and illustrated in
Fig. 2. See Fig. 8 for photographs illustrating some of
the characteristics mentioned below.
Distinguished from M. densiflorus by its much-
higher density of stem hairs, 164–235 in a 2 mm x 2
mm square near base of inflorescence where the stem
is 2–4 mm wide, vs. 26–75. Distinguished from M.
fasciculatus by its longer calyx bracts, longest ones in
inflorescence 5.5–13 mm compared to 2.5–6.0 mm,
and by its wider stipular bracts, widest ones in
inflorescence 2.0–8.0 mm, vs. 0.5–2(4) mm.
Shrub to 2.1 m tall and 2.4 m wide, generally
slightly wider than tall and approximately hemi-
spheric in shape; occasionally height 2–33width,
especially in young plants and those with lowest
branches herbivorized; mostly branched from near
base without clear central stem; occasionally with
clear central stem, but still with most larger branches
from near base; very rarely rhizomatous.
All stems ending in an inflorescence, up to ~200 in
larger plants; stem pubescence densely overlapping,
tufted to appressed stellate hairy, often becoming
sparser towards distal end of inflorescence; hairs
164–235 per 2 mm x 2 mm square sampled near base
of inflorescence where 2–4 mm wide; stem hairs
mostly without stipe, occasionally sparsely short-
stiped (0.1 mm), especially distally; stem hair rays 7–
24 averaging ~14, rays 0.1–0.2 mm averaging 0.1 mm
(rarely a few to 0.5 mm).
Leaves drought deciduous, largest occurring prox-
imally and decreasing in size distally, blade up to 71
mm long and 70 mm wide, abaxial surface paler than
adaxial surface due to trichome density; petiole to 41
mm, densely pubescent; leaves orbicular, cordate,
and/or deltoid in shape, entire or shallowly to
moderately 3–5 lobed, generally entire/wavy to
shallowly lobed, base cordate to truncate to slightly
cuneate; teeth mostly rounded to broadly-acute, 0.5–
4 mm long x 1.5–7 mm wide; rays per hair 5–27
averaging 10–15, rays 0.1–0.5 mm averaging 0.1–0.2
mm; hairs denser on abaxial side of leaf.
Inflorescence of interrupted terminal spikes with
flowers clustered in glomerules, occasionally also
with proximal axillary spikes, spikes 8–50 cm;
glomerules often densely clustered distally; terminal
inflorescence glomerules generally 4–12; glomerules
composed of 1–43 flowers in clusters of 1–5, each
flower subtended by 3 calyx bracts, each flower
cluster subtended by 1–5 stipular bracts; glomerules
often subtended by leaves and their associated
stipules, leaves in inflorescence becoming reduced
distally and often absent from more distal glomer-
ules; internodes between glomerules 5–96 mm,
generally shorter distally. See Fig. 2E for a dissected
glomerule.
Corolla width 21–42 mm; petals 10–20 mm long by
6.5–15 mm wide, asymmetrically obovate with apex
rounded, but occasionally truncate or emarginate,
pale pink to pink drying pink to pale orange/cream;
staminal column 3–7 mm; base of staminal column
and corolla lobes simple hairy to only hairy between
corolla lobes, base of staminal column occasionally
stellate hairy; blooming early April to late December
(rarely until March), peak June–August depending
on elevation. Later blooms are generally sporadic
flowers in the already fruited inflorescence. These
later blooms are possibly halted by frost damage in
the winter.
Calyx 6–15 mm long, plicate angled in bud; lobes
ovate-acuminate, length 3–10 mm, width 2–5 mm,
base width 2–4 mm; lobe widest at base to 2.5 mm
above base; lobes 1–43tube; calyx externally with
2019] 113
MORSE AND CHESTER: MALACOTHAMNUS ENIGMATICUS, A NEW RARE SPECIES
FIG. 8. Representative field photographs of M. enigmaticus. A and E. Variation in inflorescence length and compactness.
B. Typical full plant. C. Calyx with calyx bracts below. D. Immature glomerule showing stipular bracts with pair of stipules
below.
114 [Vol. 66
MADRO ˜
NO
touching to densely overlapping, tufted to appressed,
mostly sessile, stellate hairs, generally denser towards
calyx base; often with simple multicellular glandular
hairs 0.1–0.2 mm, occasionally making inflorescence
sticky; stellate hair rays 0.1–2 mm averaging 0.2–0.6
mm; calyx internally (adaxially) densely wavy hairy
on the lobes, tube and base of lobes glabrous, hairs
0.1–0.8 mm.
Calyx bracts linear, (3)5–13 mm long by 0.5–1.5
mm wide, longest generally 2/3 to .calyx; stipular
bracts ovate, linear, falcate, V-shaped, and/or Y-
shaped, 4.5–11 mm long, (0.5)1–8(12) mm wide
(width measured at widest point, which may be the
distance between the lobe tips for a lobed bract with
widely-spreading lobes), simple to 5x irregularly-
lobed distally, lobes 0.25–7 mm; stipules 4–10 mm
long by 1.5–4 mm wide; calyx bracts, stipular bracts,
and stipules tufted to appressed stellate hairy
abaxially with rays 0.1–2.3 mm averaging 0.3 mm,
adaxially simple wavy hairy similar to calyx interior.
Schizocarp pubescence of deciduous, nonstipitate,
tufted stellate hairs at distal end, otherwise glabrous;
rays 0.1–0.4 mm, longest near schizocarp center;
mericarp oval-shaped with notch at basal attach-
ment, 2.0–2.8 mm x 1.4–1.8 mm; seeds reniform, 1.4–
2.0 31–1.4 mm, white/tawny-rayed stellate hairs on
edges often reduced to golden brown tubercules on
seed faces, rays mostly ,0.1 mm, mature seed color
light to dark brown with purple, red, and/or grey
tint; seed coat generally appearing net-veined at 30x
magnification.
Paratypes: Latitude, longitude and/or elevation
that are estimated from specimen label locality
information are indicated with an asterisk (*). See
Fig. 1 for a map of localities. USA, CALIFORNIA,
San Diego Co.: Laguna Mts. Pacific Crest Trail off
Sunrise Hwy (S1), about 1 mile from parking area on
east side of hwy, 32.915, 116.453, 1700 m, 24-Dec-
2005, Bazell 1 (CHSC93162); Drainage from Storm
Canyon S-of Vallecito Valley, 32.93450, 116.41524,
665 m, 25-May-2016, Chester 2153 (TBD); Drainage
from Storm Canyon S-of Vallecito Valley, 32.93475,
116.41414, 651 m, 25-May-2016, Chester 2154
(TBD); Lower Oriflamme Canyon, 33.00949,
116.50291, 922 m, 25-May-2016, Chester 2155
(TBD); Anza Borrego Desert State Park, Culp Valley
Campground, 0.6 m NE of junction of Montezuma
Valley Rd (Highway S22) and Public Rd, at Pena
Spring, 33.2278, 116.4628, 1009 m, 12-Feb-2009,
CRES 480 (SD218779); Anza-Borrego Desert State
Park, on Pacific Crest Trail-E of Sunrise Hwy (S1),
ca-1 mi SE of its intersection with the road to Lucky
5 Ranch., 32.957, 116.504, 1567 m, 9-Jul-2005,
Determan 47 (SD193881); Laguna Mountains, 2 1/2
miles north of Shrine Camp, 32.923*, 116.476*,
1628 m*, 30-Aug-1938, Gander 6395 (SD22606);
Laguna Mountains, Desert View, 32.872*,
116.413*, 1850 m*, 31-Aug-1938, Gander 6406
(SD22617); Anza-Borrego Desert State Park-Borre-
go Palm Canyon, just east of trailhead parking at
mouth of canyon., 33.2702, 116.4148, 231 m, 16-
Mar-2006, Hendrickson 1250A (SD195245); Anza-
Borrego Desert State Park, S fork of Hellhole
Canyon, in Culp Valley, N of S 22 (Montezuma
Valley Road), 500 meters west of Pena Spring, in the
sandy wash on E facing slope, 33.226, 116.470, 1070
m, 6-Oct-2006, Manning 2 (SD193880); Laguna
Mountains, 32.91617, 116.46809, 1628 m, 18-Jun-
2015, Morse M021 (TBD); Laguna Mountains,
32.91617, 116.46809, 1628 m, 18-Jun-2015, Morse
M022 (TBD); Laguna Mountains, 32.91617,
116.46809, 1628 m, 18-Jun-2015, Morse M023
(TBD); Laguna Mountains, 32.91257, 116.46187,
1643 m, 18-Jun-2015, Morse M024 (TBD); Laguna
Mountains, 32.91257, 116.46187, 1643 m, 18-Jun-
2015, Morse M025 (TBD); Laguna Mountains,
32.91257, 116.46187, 1643 m, 18-Jun-2015, Morse
M026 (TBD); Laguna Mountains, 32.90593,
116.45620, 1658 m, 18-Jun-2015, Morse M027
(TBD); Laguna Mountains, 32.90593, 116.45620,
1646 m, 18-Jun-2015, Morse M028 (TBD); Laguna
Mountains, 32.90593, 116.45620, 1646 m, 18-Jun-
2015, Morse M029 (TBD); Mt-Laguna, 32.87471,
116.41410, 1875 m, 23-Aug-2015, Morse M040
(TBD); Mt-Laguna, 32.87553, 116.41465, 1887 m,
23-Aug-2015, Morse M041 (TBD); Anza Borrego –
Culp Valley, 33.20685, 116.48044, 1211 m, 29-May-
2016, Morse M060 (TBD); Anza Borrego – Pena
Spring, 33.22696, 116.46334, 1018 m, 29-May-2016,
Morse M061 (TBD); Anza Borrego – Culp Valley,
33.21352, 116.46206, 980 m, 29-May-2016, Morse
M062 (TBD); Anza Borrego – Culp Valley, 33.19982,
116.45392, 1000 m, 29-May-2016, Morse M064
(TBD); Anza Borrego – Culp Valley, 33.19796,
116.45299, 1045 m, 29-May-2016, Morse M065
(TBD); Anza Borrego – Culp Valley, 33.20080,
116.45296, 963 m, 29-May-2016, Morse M066
(TBD); Along PCT near Barrel Spring, south of
Montezuma Valley Rd., 33.215659, 116.58756, 1061
m, 22-Jun-2018, Morse M213 (TBD); Montezuma
Valley Road west of Ranchita, 33.218518,
116.579386, 1074 m, 22-Jun-2018, Morse M214
(TBD); Monument Peak, Laguna Mts., 32.893*,
116.420*, 1700 m, 20-Aug-1952, Munz 17955 (CAS-
BOT-BC203906, RSA78595); Laguna Mountains:
Cleveland National Forest; S of Garnet Peak and
above Storm Canyon, Penny Pines vicinity, E of
Sunrise Highway along the Pacific Crest Trail,
32.9097, 116.4550, 1685 m, 29-Jun-2005, Rebman
12129 (SD174684); Anza-Borrego Desert State Park:
Oriflamme Canyon, approx. 3.0 from route S2 on the
dirt road into the canyon, 33.0145, 116.4934, 830 m,
14-Mar-2001, Rebman 7148 (SD155153); Laguna
Mountains, near the SDSU observatory, 32.840*,
116.427*, 1860 m*, 6-Sep-1987, Reiser s.n.
(SD134433); South of Garnet Peak, 32.916*,
116.454*, 1650 m*, 17-Jul-1977, Sproul 157
(RSA388775); Pine Plantation, near Garnet Peak,
open gravelly area, 32.915*, 116.458*, 1650 m*, 15-
Jul-1977, Sproul 149 (SD114896); At edge of
chaparral among Pine Plantation South of Garnet
Peak, 32.916*, 116.454*, 1650 m*, 17-Jul-1977,
2019] 115
MORSE AND CHESTER: MALACOTHAMNUS ENIGMATICUS, A NEW RARE SPECIES
Sproul 157 (DES00029618); Anza-Borrego Desert
State Park, E of Ranchita-Culp Valley Camp Jeep
Trail, at Pena Spring, north of Highway S22,
33.2250, 116.4625, 1035 m, 19-Oct-2005, Sullivan
128 (SD182280)
EPITHET ETYMOLOGY,PRONUNCIATION,AND
COMMON NAME
The epithet refers to the 50þyear enigma
surrounding the identity of this taxon. The suggested
pronunciation is en-ig-matt’-ick-us (U.S.), or en-eeg-
matt’-eek-us (European). The suggested common
name is enigmatic bushmallow.
DISTRIBUTION AND HABITAT
Malacothamnus enigmaticus (Malvaceae) is only
known from the desert edge of the Peninsular Range
in San Diego County, from the San Ysidro Moun-
tains in the north to the Laguna Mountains in the
south.
Whether M. enigmaticus exists farther to the
south, in Baja California, is currently unknown.
Four specimens from just south of San Diego County
in Baja California, Mexico are similar to M.
enigmaticus, but are disjunct from the San Diego
County populations and do not cluster well with the
others in the PCAs. Additional specimens and
further research are needed to determine the identity
of the Baja California plants.
Malacothamnus enigmaticus is the only Malaco-
thamnus species in its geographic range, with M.
densiflorus found in the cismontane area to the west
and M. fasciculatus found in the San Jacinto / Santa
Rosa Mountains to the north. Along the ~90 km
north-south boundary between M. densiflorus and
M. enigmaticus, we found only one location with any
possible-hybrid specimens which will be investigated
further in the future. Two disjunct plants of M.
fasciculatus have been found along the creek in
Coyote Canyon northwest of Borrego Springs and
are presumably wash-downs from the Santa Rosa
Mountains. The geographic range of M. enigmaticus
and the other Malacothamnus species in the sur-
rounding area is shown in Fig. 1.
Because M. enigmaticus is a strict fire follower, it is
more difficult to pin down its distribution compared
to non-fire following species. While some sparse
plants were found to persist in sites 14 and 18 years
post-fire, in other sites the plants were all or mostly
dead just three years post-fire. The current distribu-
tion is based on historic herbarium specimens and
field surveys from 2013–2017. Future fires will likely
reveal additional plants in the distribution gaps and
possibly somewhat expand the distribution bound-
aries.
Malacothamnus enigmaticus is found almost en-
tirely on the desert side of the mountain crest. The
only specimens west of the crest are in the Laguna
Mountains, all of which are within just 0.5 km of the
crest. The main known concentrations are in the
upper alluvial fan of Cottonwood Canyon at ~900
m; the eastern slopes of Chariot Canyon at ~1200 m;
the flattish area at the Laguna Mountain Crest from
Mount Laguna to Garnet Peak at 1600–1900 m, and
in the bouldery area from just east of Ranchita to
Culp Valley at 900–1300 m. There may be popula-
tions on the steep desert slopes of the Lagunas, but if
so, they are currently unknown due to difficult
terrain and access.
There are wash-down specimens at elevations as
low as 230 m, from seeds brought there by sediment
flows after fires at higher elevations, evidenced by
specimens growing in deposited mud. Wash-down
specimens in Indian Canyon near Collins Valley
northwest of Borrego Springs indicate that there
must be a population at higher elevations in the Los
Coyotes Indian Reservation east of Hot Springs
Mountain. Unfortunately, that area has restricted
access and few voucher collections of any plant
taxon, so we cannot delineate its range there. We
have placed a question mark at this location in Fig. 1.
Malacothamnus enigmaticus occupies a unique
habitat in San Diego County, geographically sepa-
rated from similar habitat to the north and south,
where montane elevations transition rapidly to the
low desert on the east (Fig.1). This unique habitat
may have contributed to the speciation of M.
enigmaticus.
To the south of its range, montane elevations
disappear. To the north of its range, the montane
regions are not in as close proximity to low desert
elevations and are oriented differently with respect to
the low desert. The Santa Rosa Mountains are a
ridge extending into the desert, with desert on both
sides. The low desert next to the San Jacinto
Mountains is mostly in a different orientation, to
the northeast of the montane elevations instead of
almost directly east. Only M. densiflorus and M.
fasciculatus have been found in these adjacent areas.
Malacothamnus enigmaticus grows in a wide range
of vegetation communities, slopes, aspects, and soil
moistures. The only clear site preference is a recent
burn area in full or nearly-full sun. At lower
elevations, where it is warmer and drier, plants are
more common near springs, washes, and bouldery
north-facing slopes.
Due to the wide range of habitats, many
associated species were observed. The most com-
monly-observed associates include Artemisia ludo-
viciana Nutt., Bromus madritensis subsp. rubens (L.)
Husn., Bromus tectorum L., Ceanothus palmeri Trel.,
Ceanothus perplexans Trel., Cercocarpus betuloides
var. betuloides Nutt., Eriogonum fasciculatum var.
polifolium (Benth.) Torr. & A.Gray, Prunus ilicifolia
var. ilicifolia (Nutt. ex Hook. & Arn.) D.Dietr.,
Quercus cornelius-mulleri Nixon & K.P.Steele, and
Sambucus nigra subsp. caerulea (Raf.) R.Bolli.
However, these are not consistently observed at all
sites.
116 [Vol. 66
MADRO ˜
NO
CONSERVATION STATUS
Malacothamnus enigmaticus is a rare species,
known from just two areas separated by a gap of
~15 km, one in the Laguna Mountains and one in
thePinyonRidge/CulpValley/SanYsidro
Mountain area. For each of those areas, the
minimum polygonal area enclosing all known plants
has an area of 120 to 130 km
2
. The maximum area
enclosing all known plants and surrounding unex-
plored habitat that appears to be suitable for this
species is not much greater for the Laguna Moun-
tains population, but could be as high as 300 km
2
for
thePinyonRidge/CulpValley/SanYsidro
Mountain population due to the unknown extent in
the Los Coyotes Indian Reservation area.
The known populations are ~3000 plants in the
Laguna Mountains and ~500 plants in the Pinyon
Ridge / Culp Valley / San Ysidro Mountain area.
The low rainfall years after the burns in these areas
might have limited the numbers. Additional plants
are likely in areas not recently burned or with access
limitations.
Most known plants are within Anza-Borrego State
Park, Cleveland National Forest, and Bureau of
Land Management (BLM) property currently man-
aged for conservation and recreation.
Threats to the species include development, fire
suppression, chemical and/or mechanical post-fire
vegetation suppression, and possibly fires occurring
multiple years in a row which could deplete the seed
bank. All of these threats appear minimal at this
time.
Considering the limited distribution and relatively
small number of plants with a low degree of threat,
we propose a California Rare Plant Rank of 1B.3 for
M. enigmaticus. ‘‘Plants with a California Rare Plant
Rank of 1B are rare throughout their range with the
majority of them endemic to California’’ (CNPS,
2017). A threat rank of .3 indicates the taxon is ‘‘Not
very threatened in California (less than 20% of
occurrences threatened / low degree and immediacy
of threat or no current threats known)’’ (CNPS,
2017).
ACKNOWLEDGMENTS
Voucher locations from SD, RSA, and UCR, made
available online by the Consortium of California Herbaria,
were helpful in planning our fieldwork. Vouchers from SD,
RSA and UCR were the foundation of our analysis; we
thank the staff at each herbarium for providing time and
space for us to measure their Malacothamnus specimens,
and for giving us permission to remove the stem hairs from
a small area so they could be counted. We thank Carla
Hoegen and Fred Melgert for their diligence in looking for,
photographing, and recording locations of Malacothamnus
in their ~200 field surveys over the last several years in the
Laguna Mountain / Anza Borrego Desert State Park area.
Their efforts have significantly increased our knowledge of
the population sizes and distribution of Malacothamnus
species in eastern San Diego County, especially for M.
enigmaticus. We thank Bruce Watts for locations of
Malacothamnus in the San Jacinto / Santa Rosa Mountains.
We thank Elysa DuCharme for the illustrations in Fig. 2.
We thank RSA for the use of their scanning electron
micrograph to make the images in Fig. 3. We thank Jon
Rebman, J. Mark Porter, Jane Strong, Elysa DuCharme,
and Fred Melgert for reviewing and commenting on the
manuscript. We thank the many people who have accom-
panied us on field surveys, provided data via Calflora and
iNaturalist, and/or helped us in any other way. We thank
two anonymous reviewers for comments that improved this
paper.
LITERATURE CITED
BATES, D. M. 1963. The genus Malacothamnus. Unpub-
lished Ph.D. dissertation. University of California, Los
Angeles, CA.
———. 1993. Malacothamnus. Pp. 751–754 in J. Hickman
(ed.), The Jepson manual: higher plants of California.
University of California Press, Berkeley, CA.
———. 2015. Malacothamnus. Pp. 280–285 in Flora of
North America Editorial Committee, (eds.), Flora of
North America north of Mexico: Vol. 6. New York,
NY.
BEAUCHAMP, R. M. 1986. A Flora of San Diego County,
California. Sweetwater River Press, National City, CA.
BENESH,D.L.AND W. J. ELISENS. 1999. Morphological
variation in Malacothamnus fasciculatus (Torrey and
A.Gray) E.Greene (Malvaceae) and related species.
Madro ˜
no 46:142–152.
CALIFORNIA NATIVE PLANT SOCIETY (CNPS). 2017. The
CNPS ranking system. California Native Plant Society,
Sacramento, CA. Website: http://www.cnps.org/cnps/
rareplants/ranking.php [accessed 18 December 2017].
HILL, S. R. 1982. A monograph of the genus Malvastrum
A.Gray (Malvaceae: Malveae). Rhodora 84:1–83.
KEELEY, J.E., M.C. VASEY AND V.T. PARKER. 2007.
Subspecific variation in the widespread burl-forming
Arctostaphylos glandulosa, Madro ˜
no 54:42–62.
JACKSON, D. A. 1993. Stopping rules in principal compo-
nents analysis: a comparison of heuristical and
statistical approaches. Ecology 74:2204–2214.
MUNZ, P. A. 1974. A Flora of Southern California.
University of California Press, Berkeley, CA.
SLOTTA, T. A. B. 2004. Phylogenetics of the Malacothamnus
alliance (Malvaceae): assessing the role of hybridiza-
tion and molecular and morphological variation in
species delineation. Unpublished Ph.D. dissertation.
Virginia Polytechnic Institute and State University,
Blacksburg, VA.
———. 2012. Malacothamnus. Pp. 884–885 in: B. Baldwin,
D. Goldman, D. Keil, R. Patterson, T. J. Rosatti, and
D. H. Wilken (eds.), The Jepson manual: vascular
plants of California. University of California Press,
Berkeley, CA.
APPENDIX 1
KEY TO CATEGORICAL MEASUREMENTS
Number of flowers per node
1¼1–3
2¼4–9
3¼10þ
2019] 117
MORSE AND CHESTER: MALACOTHAMNUS ENIGMATICUS, A NEW RARE SPECIES
Distance from tip of calyx lobe to where the lobe is 1.0 mm
wide
1¼,1mm
2¼1–2 mm
3¼.2mm
Leaf/calyx hair density
0¼None
1¼Very sparse, could be scattered throughout, but
generally not near similar hairs
2¼Scattered throughout and near similar hairs, but most
not touching each other (exception may be the sparse but
long-rayed hairs on M. densiflorus)
3¼Most touching/overlapping, but can be easily seen
through them
4¼Densely overlapping, little can be seen through them
5¼Solid – nothing can be seen through them
Hair color presence (tawny-yellow or white)
0¼No
1¼Yes
APPENDIX 2LIST OF MEASURED CHARACTERISTICS WITH PRINCIPAL COMPONENT ANALYSIS (PCA) LOADINGS
Measured characteristic
Component Loadings for
M. enigmaticus /M. fasciculatus
Component Loadings for
M. densiflorus /M. enigmaticus
PCA1 PCA2 PCA1 PCA2
# flowers per node min 0.09 0.03 0.01 0.04
# flowers per node median 0.08 0.10 0.05 0.06
# flowers per node max 0.03 0.19 0.05 0.04
Infl branch length min (mm) 0.04 0.29 0.16 0.14
Infl branch length max (mm) 0.06 0.31 0.03 0.02
Infl internode length max (mm) 0.14 0.04 0.11 0.03
Stem thickness just below proximal infl
node (mm) 0.14 0.14 0.01 0.15
Stem thickness just above proximal infl
node (mm) 0.11 0.13 0.01 0.15
Stem thickness near distal-most infl node
(mm) 0.15 0.08 0.09 0.15
Stem stellate hairs per 2 32.3 mm
square 0.13 0.16 0.18 0.22
Stem stellate hair ray longest (mm) 0.06 0.21 0.20 0.03
Stem stellate hair ray shortest (mm) 0.01 0.02 0.09 0.18
Stem stellate hair ray average (mm) 0.11 0.23 0.20 0.18
Stem stellate hair stipe longest (mm) 0.12 0.23 0.11 0.19
Stem stellate hair stipe shortest (mm) - - 0.00 0.03
Stem stellate hair stipe average (mm) - - 0.15 0.19
Stem simple glandular hairs shortest
(mm) - - 0.14 0.02
Stem simple glandular hairs longest
(mm) - - 0.15 0.04
Stem simple glandular hairs average
(mm) - - 0.13 0.07
Stem stellate hairs tawny-yellow - - 0.03 0.05
Stem stellate hairs white 0.09 0.03 0.07 0.14
Calyx length min of 3 flowers (mm) 0.22 0.03 0.21 0.18
Calyx lobe length min of 3 flowers (mm) 0.22 0.07 0.21 0.19
Calyx lobe width min of 3 flowers (mm) 0.12 0.19 0.05 0.12
Distance of widest part of calyx lobe
from lobe base, min of 3 flowers (mm) 0.19 0.01 0.00 0.16
Distance from tip of calyx lobe to where
the lobe is 1.0 mm wide, min of 3
flowers
2
0.18 0.09 0.20 0.06
Calyx length max of 3 flowers (mm) 0.22 0.01 0.21 0.19
Calyx lobe length max of 3 flowers (mm) 0.23 0.05 0.21 0.20
Calyx lobe width max of 3 flowers (mm) 0.17 0.15 0.08 0.13
Distance of widest part of calyx lobe
from lobe base, max of 3 flowers (mm) 0.20 0.03 0.02 0.20
Distance from tip of calyx lobe to where
the lobe is 1.0 mm wide, max of 3
flowers 0.18 0.19 0.16 0.01
118 [Vol. 66
MADRO ˜
NO
APPENDIX 2CONTINUED
Measured characteristic
Component Loadings for
M. enigmaticus /M. fasciculatus
Component Loadings for
M. densiflorus /M. enigmaticus
PCA1 PCA2 PCA1 PCA2
Calyx length average of 3 flowers (mm) 0.23 0.02 0.21 0.19
Calyx lobe length average of 3 flowers
(mm) 0.23 0.07 0.21 0.20
Calyx lobe width average of 3 flowers
(mm) 0.16 0.18 0.08 0.15
Distance of widest part of calyx lobe
from lobe base, average of 3 flowers
(mm) 0.21 0.02 0.01 0.21
Distance from tip of calyx lobe to where
the lobe is 1.0 mm wide, average of 3
flowers 0.19 0.17 0.20 0.05
Calyx bract length shortest (mm) 0.22 0.11 0.20 0.15
Calyx bract length longest (mm) 0.21 0.09 0.21 0.14
Calyx bract widest part min (mm) 0.12 0.09 0.07 0.15
Calyx bract widest part max (mm) 0.14 0.19 0.12 0.14
Calyx bract margin longest hairs average
length (mm) 0.14 0.15 0.22 0.02
Calyx hair rays longest (mm) 0.15 0.21 0.23 0.06
Calyx lobe hair density 0.01 0.07 0.17 0.16
Stipular bract width min (mm) 0.20 0.05 0.12 0.21
Stipular bract width max (mm) 0.20 0.04 0.14 0.23
Distal leaf hair density upper surface 0.03 0.28 0.10 0.06
Distal leaf hair density lower surface 0.07 0.13 0.15 0.19
Proximal leaf hair density upper surface 0.01 0.17 0.08 0.04
Proximal leaf hair density lower surface 0.04 0.05 0.12 0.18
2019] 119
MORSE AND CHESTER: MALACOTHAMNUS ENIGMATICUS, A NEW RARE SPECIES
... Bract terminology follows Morse and Chester (2019). Leaf-like bracts with blades were not used in these analyses. ...
Article
Full-text available
The Malacothamnus palmeri complex (Malvaceae) includes the only three taxa in the genus with capitate to subcapitate inflorescences: M. p. var. palmeri, M. p. var. involucratus, and M. p. var. lucianus. Populations are restricted to San Luis Obispo and Monterey counties in California. No diagnostic key has been published that includes all three of these taxa. Purported complete intergradation between them has caused much confusion in identification leading to treatments combining the taxa together. These taxa occur in four regions, three of which are associated with a type specimen and the fourth of which may be the source of confusion. Here I present a morphological analysis of the M. palmeri complex to assess whether the three described taxa are morphologically distinct, whether the fourth region may include a fourth taxon, how common intermediacy is between taxa, which characters are most useful in distinguishing the taxa, and at what taxonomic rank the taxa should be recognized. I also present a key to the recognized taxa of this complex. Principal coordinate and linear discriminant analyses of morphological characters show three morphologically distinct and allopatric groups corresponding to the described varieties, with the fourth region clustering with M. p. var. involucratus. As all three taxa are both morphologically and geographically distinct with no intermediates, I elevate the varieties to the species rank with the new combinations Malacothamnus involucratus (B.L.Rob.) K.Morse and Malacothamnus lucianus (Kearney) K.Morse.
Article
Full-text available
The genus Arctostaphylos consists mostly of chaparral shrubs known by the common name manzanita, and one of the widest ranging of these is A. glandulosa Eastw., distributed from Baja California to Oregon. Particularly in the southern half of its range it exhibits complex patterns of morphological variation that have long presented taxonomic challenges. Phenetic analysis of morphological traits from over 1400 individuals from throughout the range were used to examine intra- and inter-population patterns of variation. Multivariate ordination and hierarchical cluster analysis were used to determine phenetic patterns linked with ecological and geographical distributions. These analyses suggest the hypothesis that this species comprises two lineages with a common origin but divergent in the presence or absence of glandularity: A. glandulosa Eastw. subsp. glandulosa, characterized by branchlets with long glandular hairs, scabrous or pubescent leaves, and nascent inflorescences with mostly foliaceous bracts; and A. glandulosa Eastw. subsp. cushingiana (Eastw.) Keeley, Vasey and Parker comb. nov., with non-glandular tomentose branchlets, glabrate or pubescent leaves and either foliaceous or short deltoid bracts. Populations dominated by one or the other of these morphotypes occur throughout the range and tend to be separated by elevation or distance from the coast, although mixed populations occur where these taxa come together. Two other glandular subspecies are named here. One is A. glandulosa Eastw. subsp. leucophylla Keeley, Vasey and Parker, subsp. nov., with intensely glaucous leaves and commonly with foliaceous bracts. A second glandular subspecies is A. glandulosa Eastw. subsp. atumescens Keeley, Vasey & Parker, subsp. nov., a narrowly distributed Baja California endemic similar to the nominate subspecies except that it lacks a basal burl and does not resprout after fire. Of the non-glandular tomentose taxa, in addition to A. glandulosa subsp cushingiana, several others are also recognized. One is A. glandulosa Eastw. subsp. crassifolia (Jepson) Wells, a well established coastal San Diego endemic recognized by darker and thicker leaves and smaller and flatter fruits. Another is a newly described taxon A. glandulosa Eastw. subsp. erecta Keeley, Vasey & Parker, subsp. nov., an endemic to northern Baja California recognized by the erect nascent inflorescenses. Two others have glabrate leaves and highly reduced deltoid often marcescent bracts; A. glandulosa subsp. adamsii (Munz) Wells, which has intensely glaucous leaves and is distributed from interior Riverside Co. south, and A. glandulosa Eastw. subsp. gabrielensis (Wells) Keeley, Vasey and Parker comb. nov., which has bright lustrous green leaves and greater fusion of nutlets, and is distributed from the interior San Gabriel Mountains of Los Angeles Co. north to the Sierra Madre Mountains of Santa Barbara Co. All non-glandular plants with long setose or villous hairs are A. glandulosa Eastw. subsp. mollis (Adams) Wells. This taxon includes plants with foliaceous as well as reduced bracts and is distributed throughout the Transverse Ranges from Santa Barbara to San Bernardino counties, with some outlying populations further south. This taxon shows a marked tendency for reduced stomatal densities on the upper leaf surface in the westernmost populations. Although all of the A. glandulosa taxa described here are known from allopatric populations, intergradations of these closely related taxa occur and thus some populations reflect a mixture of traits and can not be assigned a unique name of practical value.
Article
Approaches to determining the number of components to interpret from principal components analysis were compared. Heuristic procedures included: retaining components with eigenvalues (Xs) > 1 (i.e., Kaiser-Guttman criterion); components with bootstrapped Xs > 1 (bootstrapped Kaiser-Guttman); the scree plot; the broken-stick model; and components with Xs totalling to a fixed amount of the total variance. Statistical ap- proaches included: Bartlett's test of sphericity; Bartlett's test of homogeneity of the cor- relation matrix, Lawley's test of the second X; bootstrapped confidence limits on successive Xs (i.e., significant differences between Xs); and bootstrapped confidence limits on eigen- vector coefficients (i.e., coefficients that differ significantly from zero). All methods were compared using simulated data matrices of uniform correlation structure, patterned ma- trices of varying correlation structure and data sets of lake morphometry, water chemistry, and benthic invertebrate abundance. The most consistent results were obtained from the broken-stick model and a combined measure using bootstrapped Xs and associated eigen- vector coefficients. The traditional and bootstrapped Kaiser-Guttman approaches over- estimated the number of nontrivial dimensions as did the fixed-amount-of-variance model. The scree plot consistently estimated one dimension more than the number of simulated dimensions. Bartlett's test of sphericity showed inconsistent results. Both Bartlett's test of homogeneity of the correlation matrix and Lawley's test are limited to testing for only one and two dimensions, respectively.
Pacific Crest Trail off Sunrise Hwy (S1), about 1 mile from parking area on east side of hwy
  • Diego San
  • Co
San Diego Co.: Laguna Mts. Pacific Crest Trail off Sunrise Hwy (S1), about 1 mile from parking area on east side of hwy, 32.915, À116.453, 1700 m, 24-Dec-2005, Bazell 1 (CHSC93162);
Culp Valley Campground, 0.6 m NE of junction of Montezuma Valley Rd (Highway S22) and Public Rd
  • Anza Borrego Desert State Park
Anza Borrego Desert State Park, Culp Valley Campground, 0.6 m NE of junction of Montezuma Valley Rd (Highway S22) and Public Rd, at Pena Spring, 33.2278, À116.4628, 1009 m, 12-Feb-2009, CRES 480 (SD218779);
Desert State Park, on Pacific Crest Trail-E of Sunrise Hwy (S1), ca-1 mi SE of its intersection with the road to Lucky 5 Ranch
  • Anza-Borrego
Anza-Borrego Desert State Park, on Pacific Crest Trail-E of Sunrise Hwy (S1), ca-1 mi SE of its intersection with the road to Lucky 5 Ranch., 32.957, À116.504, 1567 m, 9-Jul-2005, Determan 47 (SD193881);
Desert State Park-Borrego Palm Canyon, just east of trailhead parking at mouth of canyon
  • Anza-Borrego
Anza-Borrego Desert State Park-Borrego Palm Canyon, just east of trailhead parking at mouth of canyon., 33.2702, À116.4148, 231 m, 16-Mar-2006, Hendrickson 1250A (SD195245);
Montezuma Valley Road), 500 meters west of Pena Spring
  • Anza-Borrego Desert State Park
  • Fork Of Hellhole Canyon
Anza-Borrego Desert State Park, S fork of Hellhole Canyon, in Culp Valley, N of S 22 (Montezuma Valley Road), 500 meters west of Pena Spring, in the sandy wash on E facing slope, 33.226, À116.470, 1070 m, 6-Oct-2006, Manning 2 (SD193880); Laguna Mountains, 32.91617, À116.46809, 1628 m, 18-Jun-2015, Morse M021 (TBD);