Content uploaded by Keir Morse
Author content
All content in this area was uploaded by Keir Morse on Feb 12, 2021
Content may be subject to copyright.
CROSSOSOMA
Journal of the Southern California Botanists, Inc.
Volume 44, numbers 1 & 2, pages 1-25 2018
Published January 2021
A MORPHOLOGICAL ASSESSMENT OF THE MALACOTHAMNUS
PALMERI COMPLEX (MALVACEAE)
Crossosoma 44 (1 & 2) 2018 1
A MORPHOLOGICAL ASSESSMENT OF THE MALACOTHAMNUS
PALMERI COMPLEX (MALVACEAE)
Keir Morse
California Botanic Garden
1500 N College Ave
Claremont, CA 91711
kmorse@rsagb.org
ABSTRACT
The Malacothamnus palmeri (S. Watson) Greene complex (Malvaceae) includes
the only three taxa in the genus with capitate to subcapitate inorescences: M. p.
var. palmeri, M. p. var. involucratus (R.L. Rob.) Kearney, and M. p. var. lucianus
Kearney. 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 identication 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.
KEY WORDS: Malacothamnus involucratus, Malacothamnus lucianus,
Malacothamnus palmeri, Malvaceae, morphology, new combinations, taxonomy
Crossosoma 44 (1 & 2) 2018
2
INTRODUCTION
The Malacothamnus palmeri complex includes three taxa ranked as varieties
under the species M. palmeri: M. palmeri (S. Watson) Greene var. palmeri, M.
palmeri var. involucratus (B.L. Rob.) Kearney, and M. palmeri var. lucianus
Kearney. Malacothamnus palmeri var. palmeri and M. p. var. involucratus were
originally described at the species rank, under the genus Malvastrum (Watson
1877, Robinson 1897). Malacothamnus palmeri var. involucratus was not treated
as a variety of M. palmeri until McMinn and Schumacher (1939). These three
taxa are generally treated as distinct from all other taxa of Malacothamnus
based on a capitate to subcapitate inorescence, although some plants of M. p.
var. involucratus have spike-like inorescences (Kearney 1951, Bates 1963,
Slotta 2012). The remainder of the genus have spike-like and/or panicle-like
inorescences. All three taxa are listed as rare plants by the California Native
Plant Society (CNPS) with a California Rare Plant Rank of 1B.2, which indicates
that they are moderately threatened and rare throughout their ranges (CNPS
2019). The taxonomic status of these rare plants is put into question, as recent
authors do not recognize all three taxa as distinct. Malacothamnus palmeri var.
lucianus is currently unrecognized in any treatment. The Flora of North America
(FNA) treatment (Bates 2015) does not recognize the three varieties and includes
all under M. palmeri. The current Jepson Manual treatment (Slotta 2012) includes
M. p. var. palmeri and M. p. var. involucratus, with M. p. var. lucianus subsumed
under M. p. var. palmeri. Given the conservation importance of the three named
members of the M. palmeri complex, it is critical to assess whether they are indeed
distinct from one another.
Bates (1963) notes four general regions where Malacothamnus palmeri grows in
Monterey and San Luis Obispo counties, California. Three of these are associated
with type specimens. I refer to these as the Palmeri, Lucianus, and Involucratus
regions, corresponding to the type specimen from each region. The Palmeri region
is in and around Cambria and is the only one of these regions in San Luis Obispo
County. The Lucianus region is in the Santa Lucia Mountains west of Fort Hunter
Liggett. The Involucratus region is between Lockwood and King City. The fourth
region, in and around Carmel Valley, has no type specimen associated with it but
is generally considered to have at least M. p. var. involucratus. I refer to this as
the Carmel Valley region.
Kearney (1951) wrote the rst Malacothamnus treatment to include all taxa
considered for the more recent FNA and Jepson Manual treatments with
the exception of the new taxon M. palmeri var. lucianus, which he published
four years later (Kearney 1955). In his treatment of the full genus, Kearney
distinguished M. p. var. palmeri and M. p. var. involucratus based on bract shape,
leaf pubescence and shape, and ower size, but noted they freely intergrade. His
Crossosoma 44 (1 & 2) 2018 3
description of M. p. var. lucianus notes it had recently been discovered in the
Santa Lucia Mountains, so this variety is apparently not included in his prior
note of intergradation between the other varieties. He does not provide a key to
all three taxa but describes M. p. var. lucianus as having narrower bracts than the
other two varieties, leaf pubescence and bracts approaching M. p. var. palmeri,
and the usually distinct cordate leaf bases approaching M. p. var. involucratus.
This ambiguous description is likely a reason why M. p. var. lucianus has not been
included in subsequent treatments.
In an unpublished dissertation on the genus, Bates (1963) combined all taxa
currently in the genus Malacothamnus under the species M. fasciculatus (Nuttall)
Greene and subsumed all three varieties of M. palmeri under the unpublished
name M. fasciculatus subsp. palmeri. In his published treatments (Bates 1993,
2015), he recognized M. palmeri, but not its three varieties, treating them as
synonyms of the species. In his 1963 dissertation, he notes:
“Malacothamnus fasciculatus palmeri varies considerably in the
distribution of pubescence, kinds of hairs, and the conformation of
calyces, bracts, leaves, and inorescences. With the exception of
bract and pubescence characters, however, there does not appear to
be any geographic signicance to these variations. Furthermore, all
characters show complete intergradation between their most extreme
expressions.”
Another way to interpret this is that there is signicant geographic correlation
with regard to variation in bract and pubescence characters. While Bates notes
complete intergradation in all characters, he fails to demonstrate this with any
analyses.
The geographic signicance of characters Bates referred to corresponds with the
geography of the three described varieties of Malacothamnus palmeri. He notes
two general forms of bracts below the lowermost cymules. Those from both the
Palmeri and Lucianus regions have a recognizable reduced leaf blade with two
long and broad stipules grading into often laciniate bracts with long and narrow
stipules distally. Those from the Involucratus and Carmel Valley regions have the
blade of the bract vestigial or not evident with the stipules of the bracts widened
and fused.
According to Bates, the trichomes from the Palmeri, Involucratus, and Carmel
Valley regions are similar with the exception of the leaves, which are adaxially
glabrous to glabrate in the Involucratus and Carmel Valley regions but have dense
stellate trichomes adaxially in the Palmeri region. He found those plants from the
Lucianus region to have both stellate and multicellular, simple, glandular trichomes.
He does not note these glandular trichomes in the other geographic regions. He
Crossosoma 44 (1 & 2) 2018
4
also notes longer rays in the stellate trichomes of plants from the Lucianus region.
In the Carmel Valley region, Bates notes a greater tendency for inorescences to
contain axillary cymules, leading to more spike-like inorescences. These axillary
cymules may at least partly be a function of inorescence internode length, with
capitate inorescences having highly reduced internodes and more spike-like
inorescences having longer internodes.
Slotta (2004) was the rst to undertake both molecular and morphometric analyses
of the full genus, the results of which she used for her treatment in the most
recent edition of the Jepson Manual (Slotta 2012). Her molecular analyses used
only a few loci and were unable to resolve any phylogenetic relationships within
Malacothamnus. Her morphometric analyses were more successful, and she used
them as justication to recognize some taxa relegated to synonyms by Bates.
Regarding the M. palmeri complex, she includes a UPGMA (unweighted pair
group method with arithmetic mean) dendrogram based on morphometric data
which appears to indicate that the three varieties do form three groups with M. p.
var. lucianus and M. p. var. involucratus being most similar. However, while she
did recognize M. p. var. involucratus, she treated M. p. var. lucianus as a synonym
of M. p. var. palmeri and offers no discussion on the results of her analyses related
to the M. palmeri complex or explanation for these treatment choices.
Here I present a morphological analysis of the Malacothamnus palmeri complex
to assess whether the three described taxa are morphologically distinct, whether
plants in the Carmel Valley region may be a fourth taxon, how common
intermediacy between taxa is, 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.
METHODS
Measurements
I collected morphological data on 80 specimens representing the full geographic
and morphological range of Malacothamnus palmeri, including type specimens of
all three varieties (Appendix 1). The number of specimens measured per region are
Carmel Valley (n=18), Involucratus (n=16), Lucianus (n=22), and Palmeri (n=24). I
measured duplicates of some collections to show variation within those populations,
such as the presence of both capitate and spike-like inorescences. The analyses
included 38 measurements, of which 12 were qualitative and 26 were quantitative.
Characters measured and a brief explanation of the measurements are included in
Table 1. Raw data are available at (https://doi.org/10.6084/m9.gshare.11828871).
Crossosoma 44 (1 & 2) 2018 5
Bract terminology follows Morse and Chester (2019). Leaf-like bracts with blades
were not used in these analyses. When leaf-like bracts are reduced to stipules via
the loss of the blade and petiole, I refer to these as “stipular bracts”. These can be
found in all taxa of Malacothamnus in the inorescence along the stem between
the whorl of three calyx bracts and the leaf-like bracts or leaves. Stipular bracts
are often modied in size and/or shape from more proximal stipules and in some
taxa can be fused (Figure 1). Stipular bracts from the Involucratus region are the
widest in the genus, due to both the fusion of the two adjacent modied stipules
and an increased width in each half.
A single calyx was measured per specimen. As the calyces are usually obscured in
specimens of Malacothamnus palmeri by the dense inorescence, the measured
calyx was generally either the only calyx in the fragment packet or a single calyx
removed from the inorescence. When multiple choices were available, the most
intact, well-pressed calyx was selected. A single calyx lobe per specimen was
also measured, with the same selection criteria. In general, the longest calyx bract
per calyx was measured. When the calyx was glued down, the longest accessible
calyx bract was measured.
Measurements for means of trichome length were taken from the center of the
calyx lobe measured and center of the stem internode just below the inorescence.
Ten trichomes were measured starting in the center and measuring the next closest
trichome to the previous that had not already been measured. To reduce possible
bias when choosing from the many stellate trichome rays to measure, the ruler
was held at a single angle and only rays aligned closely with this angle were
measured. Additional measurements were taken on the shortest and longest
glandular trichome and stellate trichome ray on the full calyx and the full stem
internode just below the inorescence. These measurements were intended to
represent extremes that the small sample set may not have otherwise accounted
for. For example, in some taxa, many of the longest calyx stellate trichome rays
are on the margins of the calyx and are missed by sampling the center of the lobe.
Figure 1: Representative stipular bracts. A. Lucianus region (ve bracts), B. Palmeri
region (four bracts), C. Involucratus region (one bract). Stipular bracts from the Carmel
Valley region are similar to those from the Involucratus region but generally smaller and
occasionally approaching the form of those from the Palmeri region.
Crossosoma 44 (1 & 2) 2018
6
Trichome density on the adaxial and abaxial surfaces of the leaves were quantied
using the mean distance between ten stellate trichome centers (the center point
of the trichome where the rays attach), measuring the same leaf for both sides
or leaves of similar size and maturity when glued down. A single leaf or adaxial/
abaxial leaf pair was measured per specimen. Leaf choice was focused on mature
leaves, as emerging leaves may have denser trichomes which can also be deciduous
as the leaf enlarges. Distance between trichomes was measured in a meandering
line, measuring the closest stellate trichome to the previous without remeasuring
any and moving as much as possible in a consistent direction away from the rst
trichome measured. When very sparse, all trichome distances may be measured,
as well as distance of trichomes to the nearest edge of the leaf as a substitute for
the nearest trichome when fewer than six trichomes are present. In one case where
there were no stellate trichomes on the adaxial leaf surface of a type specimen, the
width of the leaf was used as the distance between trichomes. In some sparsely
pubescent leaves with dense lines of trichomes along veins, measurements were
not allowed to follow the vein, as that would falsely increase the density estimate.
To illustrate leaf trichome density, a sample from a representative specimen from
each of the four regions was imaged using a scanning electron microscope (SEM).
Samples were mounted on aluminum stubs using conductive tape, coated with
gold using a Cressington Sputter Coater 108auto system (Cressington Scientic
Instruments, Watford, UK), and imaged on a Hitachi SU3500S Variable Pressure
SEM (Hitachi High Technologies America, Inc., Pleasanton, CA) at California
Botanic Garden.
Inorescence type was assessed on 117 specimens (Appendix 1). Each specimen
was assessed as to whether they had capitate, subcapitate, and/or spike-like
inorescences. Capitate was dened as having all owers of the inorescence in
a single capitate cluster (glomerule). Subcapitate was dened as an inorescence
with owers in a second cluster below the terminal cluster, usually with a reduced
number in the lower cluster. Spike-like was dened as an inorescence with
owers in three or more clusters. Note that a given specimen may have more than
one inorescence type.
Analyses
In order to assess clustering and intermediacy of specimens by region based on
the morphological characters measured, Principal Coordinate Analysis (PCoA)
was performed using the prcomp function in R (R Core Team, R Foundation for
Statistical Computing, Vienna, Austria). Centered and scaled PCoAs were plotted
using base functions in R. Two analyses were performed by geographic region,
one using all characters and one using all except trichome characters (see Table1).
To assess the utility of classication of regional groups based on the morphological
Crossosoma 44 (1 & 2) 2018 7
Character Explanation
In longest internode Position of longest in internode; internodes start at 1 from proximal end; 0 if in capitate
In longest internode length Longest internode in in (mm)
In longest branch length Longest in branch, measured from main stem to most distal calyx tip (mm)
In type Capitate (0), subcapitate (1), spike-like (2)
Calyx length Calyx length (to 0.5mm)
Calyx lobe length Calyx lobe length (to 0.5mm)
Calyx lobe width at widest Calyx lobe width where widest (to 0.5mm)
Calyx lobe width at base Calyx lobe width at base of calyx lobe (to 0.5mm)
Calyx lobe base to widest area Distance from center of calyx lobe base to center of where calyx lobe is widest (to 0.5mm)
Calyx lobe distance from tip where 1mm wide Distance from tip of calyx lobe to where 1mm wide (to 0.5mm)
Calyx bract length Length of longest calyx bract below measured calyx (to 0.5mm)
Calyx bract width Width of longest calyx bract (to 0.1mm)
Calyx bract color Green (0), partially red/black (1), all or mostly red/black (2)
Calyx trichomes ray length short Length of shortest calyx stellate trichome ray (to 0.1mm) from anywhere on abaxial side of calyx, including margin
Calyx trichomes ray length long Length of longest calyx stellate trichome ray (to 0.1mm) from anywhere on abaxial side of calyx, including margin
Calyx trichomes ray length mean Mean length of calyx stellate trichome rays (to 0.1mm); mean of 10 measurements from center of calyx lobe
Calyx trichomes glandular short Length of shortest calyx glandular trichomes (to 0.1mm) from anywhere on abaxial side of calyx, including margin
Calyx trichomes glandular long Length of longest calyx glandular trichomes (to 0.1mm) from anywhere on abaxial side of calyx, including margin
Calyx trichomes glandular mean Mean length of calyx glandular trichomes (to 0.1mm); mean of 10 measurements from center of calyx lobe
Calyx lobe surface visible Is calyx lobe visible through trichomes to the naked eye: no (0), slightly (1), clearly (2)
Table 1. Characters measured and explanations of measurements
Crossosoma 44 (1 & 2) 2018
8
Table 1. Characters measured and explanations of measurements
Character Explanation
Calyx tube surface visible Is calyx tube visible through trichomes to the naked eye: no (0), slightly (1), clearly (2)
Stipular bract length Width of widest stipular bract (to 0.5mm); if lobed, measured below lobes
Stipular bract width Length of widest stipular bract (to 0.5mm); if lobed, length of longest lobe
Stem trichomes ray length short Length of shortest stem stellate trichome ray (to 0.1mm) from anywhere on rst internode below in
Stem trichomes ray length long Length of longest stem stellate trichome ray (to 0.1mm) from anywhere on rst internode below in
Stem trichomes ray length mean Mean length of stem stellate trichome rays (to 0.1mm); mean of 10 measurements from center of internode
Stem trichomes glandular short Length of shortest stem glandular trichomes (to 0.1mm) from anywhere on rst internode below in
Stem trichomes glandular long Length of longest stem glandular trichomes (to 0.1mm) from anywhere on rst internode below in
Stem trichomes glandular mean Mean length of stem glandular trichomes (to 0.1mm); mean of 10 measurements from center of internode; may still be present
if rounded to 0.0mm
Stem trichomes rays visible to naked eye Are the stem stellate trichome rays visible to the naked eye; no (0), slightly (1), clearly (2)
Green stem visible to naked eye near in base Is the stem visible through the trichomes to the naked eye near the in base; no (0), slightly (1), clearly (2)
Green stem visible to naked eye distally in in Is the stem visible through the trichomes to the naked eye near the distal end of the in; no (0), slightly (1), clearly (2)
Leaf trichomes adaxial mean Mean of 10 distances between stellate trichome centers on adaxial leaf surface (to 0.1mm)
Leaf trichomes abaxial mean Mean of 10 distances between stellate trichome centers on abaxial leaf surface (to 0.1mm)
Leaf lobing Amount of leaf lobing: none (0), slight (1), obvious (2)
Leaf cordate One or more leaves with cordate base: no (0), yes (1)
Leaf truncate One or more leaves with truncate base: no (0), yes (1)
Leaf cuneate One or more leaves with cuneate base: no (0), yes (1)
Crossosoma 44 (1 & 2) 2018 9
characters measured, Linear Discriminant Analysis (LDA) was performed using
all characters (Table 1) in R using the tidyverse and MASS packages. The data
were split into two sets with 60% for training and 40% for testing. For the input
classication in the LDA, specimens were identied as one of the three varieties
of Malacothamnus palmeri using the results of the PCA. Malacothamnus palmeri
var. involucratus was further divided by the Involucratus and Carmel Valley
regions to see if the morphology in these two regions could be distinguished by
the LDA.
Box plots showing the median and the four quartiles of distribution were prepared
for select characters to show discontinuity and relative degree of overlap between
regions.
Mapping
To map the geographic distribution of Malacothamnus palmeri, I assessed the
geographic coordinates of all 246 M. palmeri specimens entered in the California
Consortium of Herbaria database (CCH 2018) and mapped them using QGIS
(QGIS Development Team 2019). Duplicate specimens were removed to allow
for only one location per collection for a total of 151 collections. I moved the
coordinates of 34 collections to more precise locations based on the location
description, retained the coordinates of 64 collections, and added coordinates
to 37 collections that did not have coordinates in CCH. To avoid misleadingly
mapped points, 16 collections with vague locations that could fall into more than
one of the M. palmeri regions and locations where plants were in cultivation were
not mapped. Including 11 additional collections I vouchered, 146 collections were
mapped and used to assess geographic isolation between taxa.
RESULTS & DISCUSSION
Analyses
The PCoA of all data shows clear separation of three groups corresponding to the
three named varieties of Malacothamnus palmeri when PC1 is plotted against
PC2, with the Involucratus and Carmel Valley regions clustering together (Figure
2). PC1 explains 43.2% of the variance and separates plants of the Lucianus
region from the rest. The character loadings of PC1 (Table 2) indicate that
the length of both glandular trichomes and stellate trichome rays are the most
inuential characters in discriminating individuals along that axis. PC2 explains
14.4% of the variance and separates the Palmeri region plants from those of the
Involucratus and Carmel Valley regions. The character loadings of PC2 establish
measurements of the calyx, leaf base shape, and density of stellate trichomes on
the leaf to be the most inuential characters on this axis. While the plants from
Crossosoma 44 (1 & 2) 2018
10
the Involucratus and Carmel Valley regions cluster together, there is some sorting
mostly along PC1. A separate PCoA (not shown) of only the Involucratus and
Carmel Valley regions shows less clear separation along a diagonal rather than
along the PC axes.
−5 0510
−2
0
2
4
6
PC 1 (43.2%)
PC 2 (14.4%)
Carmel Valley
Involucratus
Lucianus
Palmeri
Figure 2: PCoA results by geographic region using all characters measured. Type specimens
have larger symbols with bold borders. Note the lectotype and syntype are included for
Malacothamnus palmeri var. involucratus. Both specimens have an annotation calling it
the holotype and both have similar PC values, which shows either to have been a reasonable
choice for the lectotype.
As trichomes were important characters separating the three taxa on both PC1
and PC2 in the PCoA using all characters, I ran a separate PCoA excluding
all trichome characters to test separation of the three taxa based only on non-
trichome characters (Figure 3). While separation is not quite as extreme as with
all characters, the three described taxa still separate with no overlap. Separation
between the Carmel Valley and Involucratus regions is mostly lost in this analysis.
The LDA resulted in 100% correct classication of individuals pre-assigned
by region. Plants from the Carmel Valley region again were very close to the
Involucratus region relative to the others. LDA coefcients are listed in Table 2.
Crossosoma 44 (1 & 2) 2018
11
Table 2. PC loading and LDA coefcients of characters measured
Character PC1 PC2 LD1 LD2 LD3
In longest internode -0.12 -0.20 -8.93 -6.20 -1.87
In longest internode length -0.10 -0.17 0.31 -0.20 -0.05
In longest branch length 0.13 0.15 1.29 0.46 -0.14
In type -0.12 -0.19 -2.79 3.37 2.40
Calyx length 0.15 0.29 -1.03 -0.34 2.09
Calyx lobe length 0.15 0.29 -1.74 3.12 -1.46
Calyx lobe width at widest 0.01 0.34 3.02 -0.63 0.97
Calyx lobe width at base 0.00 0.23 -3.79 2.86 -1.37
Calyx lobe base to widest area 0.08 0.33 -2.32 5.90 -0.57
Calyx lobe distance from tip where 1mm wide 0.17 0.00 4.42 -2.75 -0.68
Calyx bract length 0.15 0.19 -1.55 -0.20 0.07
Calyx bract width -0.13 -0.08 5.69 -7.45 -1.91
Calyx bract color 0.11 -0.05 39.37 -4.79 -1.37
Calyx hairs ray length short 0.18 -0.03 -196.52 85.66 -23.68
Calyx hairs ray length long 0.18 0.07 7.64 3.26 0.16
Calyx hairs ray length mean 0.23 -0.08 76.05 -30.86 -10.80
Calyx hairs glandular short 0.24 -0.07 -116.12 193.97 -42.46
Calyx hairs glandular long 0.23 -0.05 157.27 -70.85 10.72
Calyx hairs glandular mean 0.24 -0.05 114.20 36.00 -31.03
Crossosoma 44 (1 & 2) 2018
12
Table 2. PC loading and LDA coefcients of characters measured
Character PC1 PC2 LD1 LD2 LD3
Calyx lobe surface visible 0.20 -0.14 0.77 2.41 1.27
Calyx tube surface visible 0.10 -0.08 1.83 -9.00 1.08
Stipular bract length 0.15 0.05 0.84 -1.24 0.25
Stipular bract width -0.15 -0.17 -1.38 -0.44 0.04
Stem hairs ray length short 0.02 0.09 -15.27 30.61 -12.78
Stem hairs ray length long 0.23 -0.03 13.10 9.39 -4.26
Stem hairs ray length mean 0.22 0.00 3.35 -24.94 -3.89
Stem hairs glandular short 0.23 -0.08 10.93 81.61 60.85
Stem hairs glandular long 0.23 -0.08 -52.29 -12.87 32.37
Stem hairs glandular mean 0.23 -0.09 -35.76 60.09 -50.06
Stem hairs rays visible to naked eye 0.22 -0.08 59.78 -20.08 11.81
Green stem visible to naked eye near in base 0.19 -0.16 16.50 -18.73 -1.01
Green stem visible to naked eye distally in in 0.19 -0.16 -25.55 8.96 2.09
Leaf hairs adaxial mean -0.05 -0.11 0.78 -0.62 0.35
Leaf hairs abaxial mean 0.03 -0.24 45.43 -68.64 -12.66
Leaf lobing -0.09 -0.10 10.19 0.02 -2.60
Leaf cordate 0.06 -0.27 3.90 -4.95 3.18
Leaf truncate -0.12 0.13 -6.10 -1.60 -1.37
Leaf cuneate -0.10 0.20 -3.62 6.64 -0.39
Crossosoma 44 (1 & 2) 2018
13
Figure 3: PCoA results by geographic region without trichome characters. Type specimens
have larger symbols with bold borders. Note the lectotype and syntype are included for
Malacothamnus palmeri var. involucratus; see Figure 2 caption for more explanation.
−6 −4 −2 024 6
−4
−2
0
2
PC 1 (35.7%)
PC 2 (15.1%)
Carmel Valley
Involucratus
Lucianus
Palmeri
The PCoAs demonstrate that plants from the Involucratus, Lucianus, and Palmeri
regions are morphologically distinct from one another. These regions correspond
to the described varieties of Malacothamnus palmeri with the Carmel Valley
and Involucratus regions clustering together. Separation of these three varieties
based on the characters examined is very clear. The PCoAs and LDAs indicate
many useful characters in distinguishing the three varieties when considered as
a whole, however many measurements overlap between these varieties. I found
three groupings of characters that were the most consistent and easy to use for
diagnosis of taxa: trichome length, leaf trichome density, and stipular bract width
and shape. These are examined below relative to the four regions.
Trichome length
All specimens examined had glandular trichomes on the stems, leaves, and
inorescences. Those from the Lucianus region were much larger than those from
the other regions, ranging from 0.1-1.4 mm with most >=0.3 mm (mean 0.5 mm).
In the remaining regions, these trichomes were considerably reduced (<=0.1 mm)
and were often only detectable as a drop of exudate. Glandular trichomes less
than 0.1 mm in length were recorded as 0 mm long. Exudate from the glandular
trichomes was often easier to see than the trichomes themselves and ranged in
Crossosoma 44 (1 & 2) 2018
14
color from dark reds to yellows as well as clear. Glandular trichomes concealed
within this exudate were conrmed with scanning electron microscopy (Figure 4).
The rays of the stellate trichomes were likewise much larger in the Lucianus
region and particularly distinct on the stem, with rays ranging from 0.1-4.8 mm
with many >1 mm on all plants (mean 1.2 mm). In contrast, those from the other
regions measured 0.1-1.3 mm with most <1 mm on all plants (mean 0.4 mm).
Boxplots of the longest stellate trichome rays and glandular trichomes on the stem
clearly illustrate how those from the Lucianus region can be easily distinguished
from the rest based on length of trichomes (Figures 5 and 6).
Leaf trichome density
Density of stellate trichomes on the adaxial surface of mature leaves showed
clear patterns by region. Those from the Involucratus region are sparsest, with
the centers of the trichomes averaging 1.3 mm apart on the densest specimen
to glabrous on the sparsest. Specimens from the Carmel Valley region overlap
with the Involucratus region, with trichomes averaging 0.4-4 mm apart. With the
exception of a single outlier specimen, plants from the Palmeri region have the
densest trichomes, with trichomes averaging 0.16-0.26 mm apart. Those from the
Lucianus region overlap those of the Carmel Valley and Palmeri regions, with
trichomes averaging 0.26-0.69 mm apart. While the density of trichomes of the
Lucianus region specimens are intermediate to those of the other regions, the
Lucianus plants can be separated based on length of trichomes, which leaves a
clear gap between the density of trichomes in the Palmeri region compared to the
Involucratus and Carmel Valley regions for most specimens (Figure 7). Scanning
electron micrographs of representative adaxial leaf trichome density for each
region are shown in Figure 8.
Figure 4: Scanning electron micrograph showing simple glandular trichomes on the
stem of Malacothamnus palmeri var. palmeri, the largest of which would be rounded in
measurement to 0.1 mm and the others to 0 mm. Note the bases of broken off stellate
trichomes resembling stumps of trees.
Crossosoma 44 (1 & 2) 2018 15
Figure 5: Boxplots of length of the longest glandular trichomes on the stem by
geographic region.
Figure 6: Boxplots of length of the longest stellate trichome rays on the stem by
geographic region.
Carmel Valley
n=18
Involucratus
n=16
Lucianus
n=22
Palmeri
n=24
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Longest glandular trichomes (mm)
Carmel Valley
n=18
Involucratus
n=16
Lucianus
n=22
Palmeri
n=24
1234
Longest stellate trichome rays (mm)
Crossosoma 44 (1 & 2) 2018
16
Stipular bracts
The range of stipular bract widths from each region overlaps with at least one other
region (Figure 9), but there are patterns that can generally distinguish between
regions. Stipular bract width in the Involucratus region for all specimens was
greater than both the Palmeri and Lucianus regions but had considerable overlap
with the Carmel Valley region. When the Involucratus and Carmel Valley regions
are considered together, 94% of specimens had a width >= 7 mm, compared to
96% from the Palmeri region having a width <=6.5 mm. Stipular bract width in
the Palmeri region is generally larger than the Lucianus region, with 75% from
the Lucianus region and only 4% from the Palmeri region having bracts <= 2 mm
wide.
The shape of the stipular bracts is also signicant between regions, as seen by
the length to width ratio (Figure 1 and 10). In the Involucratus and Carmel Valley
regions, the stipular bracts are generally about as wide as long and two-lobed due
to the fusing of two modied stipules. In the Palmeri region, they are generally
linear to lanceolate and unlobed. In the Lucianus region, they are generally linear
and unlobed.
Carmel Valley
n=18
Involucratus
n=15
Lucianus
n=22
Palmeri
n=24
0510 15
Mean distance between stellate trichomes (mm)
Figure 7: Boxplots of mean distance between the central axis of stellate trichomes on
the adaxial leaf surface by geographic region, indicative of stellate trichome density. One
specimen from the Involucratus region had no stellate trichomes on the adaxial leaf surface
and is not included here.
Crossosoma 44 (1 & 2) 2018 17
Figure 8: Scanning electron micrographs showing representative stellate trichome density
on the adaxial leaf surface for each geographic region of Malacothamnus palmeri - A.
Involucratus, B. Carmel Valley, C. Lucianus, and D. Palmeri.
Carmel Valley
n=18
Involucratus
n=16
Lucianus
n=22
Palmeri
n=24
0 5 10 15 20 25
Widest stipular bract width (mm)
Figure 9: Boxplots of width of widest stipular bract by geographic region.
Crossosoma 44 (1 & 2) 2018
18
Inorescence type
A larger set of specimens than the group used for the previous analyses was
assessed for inorescence type. Data collection allowed for multiple inorescence
types to be recorded for the same specimen. Multiple inorescence types were
found on the same specimens both from multiple stems and from single stems
with branches. Of the 117 specimens examined for inorescence type, 102 had
a capitate inorescence, 48 had a subcapitate inorescence, and 113 had capitate
and/or subcapitate inorescences. Only four specimens did not have a capitate or
subcapitate inorescence. Most specimens (97%) in the Involucratus, Lucianus,
and Palmeri regions had capitate inorescences. Only 63% of specimens from
the Carmel Valley region had capitate inorescences. Subcapitate inorescences
were rare (12%) in the Lucianus and Palmeri regions, but 63% of specimens in
the Carmel Valley region and 38% of specimens in the Involucratus region were
subcapitate.
Having a spike-like inorescence is somewhat problematic for identication in
the Malacothamnus palmeri complex as the complex is generally separated from
the rest of the genus based on having a capitate to subcapitate inorescence. Of
the 117 specimens examined, 18 had spike-like inorescences. All of these were
from the Involucratus and Carmel Valley regions. Of the 18 specimens with spike-
like inorescences, 14 also had capitate and/or subcapitate inorescences. The
remaining four specimens possessing only spike-like inorescences are from
Figure 10: Boxplots of ratio of the length to width of stipular bracts by geographic region.
Carmel Valley
n=18
Involucratus
n=16
Lucianus
n=22
Palmeri
n=24
0 5 10 15 20 25 30
Ratio of stipular bract length to width
Crossosoma 44 (1 & 2) 2018 19
three locations. Two of these locations have additional specimens that have either
capitate or subcapitate inorescences. This leaves a single specimen location near
Pacic Grove (CAS363409) without capitate or subcapitate inorescences. As
all other collected populations had some capitate or subcapitate inorescences,
it is likely that all or most populations with spike-like inorescences also had
plants with capitate or subcapitate inorescences, and thus using a capitate to
subcapitate inorescence as a diagnostic trait for identication should still work if
the population as a whole is considered. How common spike-like inorescences
are relative to capitate or subcapitate inorescences in these regions is unknown.
In surveys over the past three years, I was only able to nd eight plants in three
populations from the Involucratus and Carmel Valley regions. As Malacothamnus
are re-followers and generally short-lived, variation in inorescence types would
be best assessed when future res allow new plants to germinate (Bates 1963).
Bates’ published treatments (1993, 2015) possibly tried to account for some
Malacothamnus palmeri specimens having spike-like inorescences by using
a different set of characters to identify M. palmeri. The rst couplet of these
treatments, however, key 76% of M. p. var. lucianus and 25% of M. p. var. palmeri
specimens measured here to taxa outside the M. palmeri complex, and those with
spike-like inorescences still fall in the ambiguous couplet leading to M. palmeri.
Using the presence of capitate to subcapitate inorescences as the rst couplet
reduces the error in keying to the M. palmeri complex to only 3% of specimens
examined and <1% if the variation in duplicate specimens is considered.
Intergradation
No clear morphologically intermediate specimens were observed between the
Lucianus, Involucratus, and Palmeri regions. The complete intergradation of
characters suggested by Bates may be entirely related to plants from the Carmel
Valley region. Many of these plants have denser trichomes than typical for the
Involucratus region, but not as dense as the Palmeri and Lucianus regions. Several
Carmel Valley specimens are ambiguous in the character of wide and fused stipular
bracts. These may be obscured on the specimens, missing, or of intermediate
form. Considering those two characters, along with the fact that many plants
from the Carmel Valley region have spike-like inorescences which are not found
in the Lucianus and Palmeri regions, it seems more likely that Malacothamnus
palmeri var. involucratus is more variable than previously dened and/or these
aberrant plants have some genetic inuence from another taxon with a spike-like
inorescence outside of the M. palmeri complex. Despite the somewhat confusing
variation in the Carmel Valley plants, they clearly cluster morphologically with
the Involucratus plants and would be difcult to reliably distinguish from them.
Crossosoma 44 (1 & 2) 2018
20
Range
Measured specimens were mapped by variety following the PCoA results, along
with additional specimens keyed based on characters found most useful in the PCoA
results (Figure 11). Excluding a disjunct specimen of Malacothamnus palmeri
var. involucratus discussed below, M. p var. involucratus occurs exclusively in
the Involucratus and Carmel Valley regions, M. p. var palmeri occurs exclusively
in the Palmeri region, and M. p. var. lucianus occurs exclusively in the Lucianus
region. Malacothamnus palmeri var. involucratus has a gap of ~53 km between
its two regions, a gap of ~20 km to the nearest location of M. p. var. lucianus, and
a gap of ~45 km to the nearest location of M. p. var. palmeri. There is a gap of
~70 km between the nearest M. p. var. lucianus and M. p. var. palmeri locations.
A single disjunct specimen of Malacothamnus palmeri var. involucratus (C.
Dudley s.n. 1 May 1937 CAS) was purported to be found at Cuesta Pass ~15 km
southeast of the closest M. p. var. palmeri location and ~95 km south-southeast
of the closest location of M. p. var. involucratus (Figure 11). This specimen was
possibly found adjacent to a major road and may be from a human introduction.
Alternatively, this location is in line with the two M. p. var. involucratus regions
and could be part of a larger historic range. A third possibility is that the location
on the label is erroneous. All other collections by Dudley from that same day are
attributed to “S.W. part of Monterey Co.,” while this disjunct location is in San
Luis Obispo County. No Malacothamnus of any taxon were found in 2018 and
2019 surveys in the Cuesta Pass area.
While the geographic gaps between these three taxa are relatively small, surveys in
these gaps have yielded only very minor range extensions, and the gaps are likely
large enough to prevent cross-pollination between taxa. Studies on foraging ranges
of bees have found solitary bees with a maximum foraging distance of 1.4 km and
honey bees foraging as far as 14.5 km from their hive (Beekman and Ratnieks
2000, Zurbuchen et al. 2010). This is less than the closest geographic gap between
Malacothamnus palmeri varieties of ~20 km except for the questionable Cuesta
Pass location. Garden experiments by Bates (1963) found many Malacothamnus
taxa could cross and form morphologically intermediate hybrids, but the lack of
any morphologically intermediate specimens between these three taxa suggests
that the geographic isolation between these regions is an effective reproductive
barrier, even if they are capable of crossing in articial situations.
CONCLUSIONS
The unied species concept (de Queiroz, 2007) denes species based on the
common element of previous species concepts, that species are separately-evolving
metapopulation lineages. This concept considers additional criteria used to dene
Crossosoma 44 (1 & 2) 2018 21
Figure 11: Range of Malacothamnus palmeri mapped by variety with both assessed and
unassessed specimens shown. Assessed specimens include both measured and keyed
specimens. Unassessed specimens include all other specimens on the CCH with a mappable
location that had been identied to Malacothamnus palmeri (or synonyms) at the time of
this study, but which have not yet been examined by the author. Type specimens have a
larger icon. Both syntypes of M. palmeri var. involucratus are shown. Inset shows extent
of map within California.
Crossosoma 44 (1 & 2) 2018
22
species in previous concepts as lines of evidence for this denition, recognizing
that not all lineages will meet all the criteria or meet these criteria in the same
chronological order. Results of both the PCoA and LDA support three distinct
taxa based on morphology, which correspond to the three described varieties of
Malacothamnus palmeri. All specimens assessed were easily assigned to one of
the three taxa with no morphologically intermediate specimens observed. Each
taxon is allopatric, limiting any potential gene ow between them. Preliminary
maximum likelihood analyses of DNA sequences show 100% bootstrap support
for each taxon belonging to their own clade (Morse, unpublished data). Based on
these lines of evidence, it is clear that the three described varieties of M. palmeri
are separately-evolving metapopulation lineages and thus t the species denition
of the unied species concept. Hence, I elevate M. p. var. involucratus and M. p.
var. lucianus to the species rank as:
Malacothamnus involucratus (B.L.Rob.) K.Morse, comb. nov. Basionym:
Malvastrum involucratum B.L.Rob. in A.Gray, Syn. Fl. N. Amer. 1(1): 310. 1897.
Lectotype designated by Kearney (1951): USA, California, Monterey County,
Jolon, date unknown, T. S. Brandegee s.n. (GH52904)
Malacothamnus lucianus (Kearney) K.Morse, comb. et stat. nov. Basionym:
Malacothamnus palmeri (S.Watson) Greene var. lucianus Kearney, Lea. W. Bot.
7: 289-290. 1955. Holotype: USA, California, Monterey County, Santa Lucia
Mts., Arroyo Seco road 3 miles from Escondido Camp, 2600 ft. elevation, 8 July
1955, John Thomas Howell 30642 (CAS396251)
While the LDA could separate the Carmel Valley plants from the rest of
Malacothamnus involucratus 100% of the time for those specimens analyzed,
reliably separating them without such an analysis seems impractical at present,
especially for eld identication. The PCoA mostly separated these plants but
was less clear. It could be argued that the Carmel Valley plants should be treated
as a variety of M. involucratus, but it may be best to consider them a divergent
population until there is further evidence.
Crossosoma 44 (1 & 2) 2018 23
Key to the species in the Malacothamnus palmeri complex
The following key separates all three species in the Malacothamnus palmeri
complex with measurements based on the dry specimens used in the analyses
for this paper. The key was tested on an additional 91 specimens not used in the
analyses with 100% accuracy in species determination. Appendix 1 includes a list
of all specimens examined for the analyses and the key.
1. Many rays of stellate trichomes on stem 1-3 mm; many simple glandular
trichomes 0.3-1.4 mm, generally distinct at 20x magnication, occasionally
sparse and difcult to detect; leaves often with a rancid odor; surface of stem and
calyx lobes generally easily visible through trichomes without magnication - M.
lucianus
1’. Most rays of stellate trichomes on stem <1 mm; glandular trichomes <=0.1
mm, often only apparent as a resinous dot, much smaller than and often obscured
by adjacent stellate trichomes; surface of stem and calyx lobes often hidden by
dense trichomes - 2
2. Adaxial leaf surface with dense stellate trichomes in mature leaves, centers of
stellate trichomes average <=0.25 mm apart, rays of adjacent trichomes generally
overlapping across entire leaf surface; inorescence with stipular bracts linear to
lanceolate and unlobed, widest stipular bracts <=6.5(9) mm wide - M. palmeri
2’. Adaxial leaf surface glabrous or with sparse stellate trichomes in mature
leaves, if trichomes more dense, centers of stellate trichomes average >=0.5
mm apart, rays of adjacent trichomes not overlapping across entire leaf surface;
inorescence generally with lobed stipular bracts that are about as wide as long,
widest stipular bracts >=7(5) mm wide measured below lobes - M. involucratus
ACKNOWLEDGMENTS
Thanks to J. Mark Porter for advice throughout this study and for useful
comments and suggestions on the manuscript. Thanks to reviewers Dieter Wilken,
Marc Baker, Michael Simpson, and editor/reviewer Michelle Cloud-Hughes for
corrections and suggestions that greatly improved the manuscript. Thanks to Amy
Patten, Angelique Herman, Dave Nelson, Mike Splain, Nikki Nedeff, and Vern
Yadon who helped with surveys and/or information related to the Malacothamnus
palmeri complex. Thanks to CAS/DS, CDA, DAV, GH/A, OBI, SBBG, and SJSU
herbaria for loaning voucher specimens for this study and the staff at RSA/POM
for facilitating these loans. Thanks to Calora and iNaturalist for facilitating the
mapping and sharing of plant observations. Thanks to the following organizations
and people for funding this study and the larger study of the full genus currently
Crossosoma 44 (1 & 2) 2018
24
in progress: American Society for Plant Taxonomists; Anza Borrego Foundation;
Arizona Native Plant Society; California Botanic Garden; California Native Plant
Society, including the Bristlecone, Orange County, San Gabriel Mountains, Santa
Clara Valley, and Shasta Chapters; Claremont Graduate University; Michael
Hagebusch; Northern California Botanists; and Southern California Botanists.
Likewise, thanks to the multitude of people who have helped in many ways on the
larger study in progress of which this is but a small part.
LITERATURE CITED
Bates, D.M. 1963. The genus Malacothamnus. Ph.D. dissertation. University of
California, Los Angeles, CA. 171 pp.
––––––. 1993. Malacothamnus. Pages 751-754 in J.C. Hickman (ed.), The Jepson
Manual: Higher Plants of California. Berkeley: University of California
Press, Berkeley.
––––––. 2015. Malacothamnus. Pages 280-285 in Flora of North America
Editorial Committee (eds.), Flora of North America North of Mexico. Vol.
6. Oxford University Press. New York and Oxford.
Beekman, M. and F.L.W. Ratnieks. 2000. Long-range foraging by the honey bee,
Apis mellifera L. Functional Ecology 14:490-496.
California Native Plant Society (CNPS). 2019. CNPS Rare Plant Ranks. Online:
https://www.cnps.org/rare-plants/cnps-rare-plant-ranks (accessed 21 Nov
2019).
Consortium of California Herbaria (CCH). 2018. Online: http://ucjeps.berkeley.
edu/consortium/ (accessed 9 Feb 2018).
De Queiroz, K. 2007. Species Concepts and Species Delimitation. Systematic
Biology 56:879-886.
Kearney, T.H. 1951. The Genus Malacothamnus, Greene (Malvaceae). Leaets of
Western Botany VI:113-140.
––––––.1955. Notes on Malvaceae VII: A New Variety in Malacothamnus.
Leaets of Western Botany VII:289-290.
McMinn, H. and F.H. Schumacher. 1939. An Illustrated Manual of California
Shrubs. J. W. Stacey, Incorporated, San Francisco.
Morse, K. and T. Chester. 2019. Malacothamnus enigmaticus (Malvaceae), a new
rare species from the desert edge of the Peninsular Range in San Diego
County, CA. Madroño 66:103-119.
QGIS Development Team. 2019. QGIS 2.4.4-Madiera. QGIS Geographic
Information System. Open Source Geospatial Foundation Project. https://
qgis.org
Robinson, B.L. 1897. Malvastrum. Pages 308-313 in B.L. Robinson (ed.),
Synoptical Flora of North America. Vol.1 Part 1. American Book Company,
New York.
Crossosoma 44 (1 & 2) 2018 25
Slotta, T. 2004. Phylogenetics of the Malacothamnus alliance (Malvaceae):
assessing the role of hybridization and molecular and morphological
variation in species delineation. Ph.D. dissertation. Virginia Polytechnic
Institute and State University, Blacksburg, Virginia. 236 pp.
––––––. 2012. Malacothamnus. Pages 884-885 in B. Baldwin, D. Goldman, D.
Keil, R. Patterson, T.J. Rosatti, D.H. Wilken, (eds.), The Jepson Manual:
Vascular Plants of California. University of California Press, Berkeley.
Watson, S. 1877. Descriptions of new species of plants, with revisions of certain
genera. Proceedings of the American Academy of Arts and Sciences
12:246-278.
Zurbuchen, A., Landert, L., Klaiber, J., Müller, A., Hein, S., and Dorn, S. 2010.
Maximum foraging ranges in solitary bees: only few individuals have
the capability to cover long foraging distances. Biological Conservation
143:669-676.