ArticlePDF Available

Geographic variation in flower color patterns within Calceolaria uniflora Lam. in Southern Patagonia

Authors:
  • National Scientific and Technical Research Council, Córdoba

Abstract and Figures

Infraspecific variation in flower colors was evaluated in 26 populations of Calceolaria uniflora Lam. in Southern Patagonia, Argentina. Computerized analysis of high-resolution photo-images was used to estimate the proportions of red, orange and yellow in surfaces of two corolla parts, instep and throat, in field samples of 20–35 flowers per population. The between-populations component accounted for 48% of variance for instep colors and 24% for throat colors. Geographic differentiation was found between populations with a uniform red instep in the Andes in the west, and populations with a maculate yellow-and-red instep in the Magellanic steppe to the east. Mixed populations occurred in a transition zone. Throat colors showed a different, north-south geographic trend. Based on color pattern and distribution, two subspecies may be differentiated within C. uniflora. Their overall geographic distribution is related to climate and vegetation, but their detailed distribution is better explained by isolation by distance and barriers to gene flow.
Content may be subject to copyright.
Geographic variation in flower color patterns within Calceolaria
uniflora Lam. in Southern Patagonia
M. Masco
´
1
, I. Noy-Meir
2
,and A. N. Se
´rsic
3
1
EEA Santa Cruz, Consejo Agrario Provincial, INTA & Universidad Nacional de la Patagonia Austral,
´o Gallegos, Argentina
2
Institute of Plant Sciences, Faculty of Agricultural, Food and Environmental Quality Sciences, Hebrew
University of Jerusalem, Israel
3
Instituto Multidisciplinario de Biologı
´a Vegetal, Universidad Nacional de Co
´rdoba & CONICET,
Argentina
Received March 3, 2003; accepted September 7, 2003
Published online: February 3, 2004
ÓSpringer-Verlag 2004
Abstract. Infraspecific variation in flower colors
was evaluated in 26 populations of Calceolaria
uniflora Lam. in Southern Patagonia, Argentina.
Computerized analysis of high-resolution photo-
images was used to estimate the proportions of
red, orange and yellow in surfaces of two corolla
parts, ‘‘instep’’ and ‘‘throat’’, in field samples of
20–35 flowers per population. The between-popu-
lations component accounted for 48% of variance
for instep colors and 24% for throat colors.
Geographic differentiation was found between
populations with a uniform red instep in the
Andes in the west, and populations with a
maculate yellow-and-red instep in the Magellanic
steppe to the east. Mixed populations occurred in
a transition zone. Throat colors showed a differ-
ent, north-south geographic trend. Based on color
pattern and distribution, two subspecies may be
differentiated within C. uniflora. Their overall
geographic distribution is related to climate and
vegetation, but their detailed distribution is better
explained by isolation by distance and barriers to
gene flow.
Key words: Infraspecific variation, populations,
image analysis, multivariate analysis, Argentina.
Introduction
The patterns of variation in characters between
populations within a taxon and in related taxa
are the subject matter of biosystematics (Solbrig
1970, Elkington 1986, Stace 1989). Variation
between geographically separated populations
within a species is of special interest as an indi-
cation of divergence and incipient speciation
(Judd et al. 1999). Most studies of infraspecific
variation at a geographic scale have been
conducted in Europe and North America.
There are few such studies on plants of
temperate-cold regions in the southern
hemisphere, such as Southern Patagonia.
Biosystematic studies of angiosperm taxa
often focus on morphological characters of the
flower (shape, color) because they usually are
phenotypically more stable than vegetative
characters and vary less within populations
than molecular markers (e.g. Podolsky and
Holtsford 1995). Variation in flower shape and
color has commonly been interpreted as
‘‘adaptive’’, resulting from disruptive selection
Plant Syst. Evol. 244: 77–91 (2004)
DOI 10.1007/s00606-003-0083-1
by different pollinators (Darwin 1859, Faegri
and van der Pijl 1979, Schemske and Bradshaw
1999, Clegg and Durbin 2000). However, other
processes can produce floral divergence be-
tween geographically segregated populations,
e.g. random drift and isolation by distance,
indirect selection and genetic context (Judd
et al. 1999, Schemske and Bierzychudek 2001,
Armbruster 2002).
The research presented here was designed
as a biosystematic study of patterns of infra-
specific variation in floral colors at a geo-
graphical scale in a subantarctic species of
Calceolaria (Scrophulariaceae). Pollination
biology and selection by pollinators were not
directly studied but are considered in the
discussion among the possible mechanisms
that may have generated the observed patterns
of variation.
Reflectance spectrophotometry is com-
monly used to quantify flower colors, applying
probes to areas of uniform color (e.g. Mele
´n-
dez-Ackerman 1997). Fine-grained spatial pat-
terns of two or more colors, as in flowers of
Calceolaria (and many other species), are not
easily quantified. In taxonomy, floriculture
and plant breeding, descriptive terms (e.g.
spotted, striated) are used. Semi-quantitative
scores can be obtained by visual estimates of
the proportion of area covered by patches of
different colors. However, such estimates are
inconsistent where the patches are small and
fuzzy (personal observations). An objective
method to measure the proportion of patches
of different colors, by computerized analysis of
photoimages of flowers, was developed here in
the biosystematic context of infraspecific var-
iation in Calceolaria. This technique does not
detect patterns of colors invisible to humans
but visible to some pollinators (such as UV)
and therefore the results presented here do not
necessarily express flower color as perceived by
pollinators (Chittka et al. 1994, Vorobyev
et al. 1997).
The objective of this biosystematic re-
search was to analyze quantitatively the
variation in flower colors within Calceolaria
uniflora, in relation to geographic and
ecological gradients. The following questions
were addressed:
Can distinct types of floral color pattern be
visually identified within the species?
What is the relative magnitude of between-
populations and within-population variation
in the proportion of flower areas covered by
different colors?
To what extent is this variation either
phenotypically stable, or plastic in response
to environmental conditions?
Are colors in different parts of the flower
correlated or independent?
Are there clear trends in flower colors in
geographic space and in relation to climatic
variables?
Are there well-defined clusters of popula-
tions in geographic and character space, so
that infraspecific taxa may be identified? Or is
variation between populations entirely clinal?
We will finally attempt to interpret the
observed patterns of geographic variation
within the species in relation to alternative
evolutionary processes that may have gener-
ated them.
Materials and methods
Calceolaria uniflora Lam. The genus Calceolaria
contains about 270 species in South America,
mostly in the Andean region (Molau 1988).
C. uniflora has a subantarctic distribution at the
southernmost limit of the genus, in Southern
Patagonia and Tierra del Fuego (in both Argen-
tina and Chile) between latitudes 45°and 55°S
(Descole and Borsini 1954, Moore 1983, Boelcke
et al. 1985, Masco
´et al. 1998, Roig 1998, Correa
1999, Ehrhart 2000). It is found in a variety of
habitats from sea level on the Atlantic coast and
the Magellan Strait to over 1200 m a.s.l. in the
Andes Mountains.
The characteristic Calceolaria corolla includes
a small upper lip enclosing the fertile parts, and a
lower lip enlarged in the form of a shoe or slipper
(Fig. 1). The surfaces of the lower lip frontally
exposed to pollinators are the ‘‘throat’’ (the open
basal part) and the ‘‘instep’’ (the upper surface of
the saccate distal part). In most species, the
proximal edge of the instep forms a lap that is
folded inward, hiding the oil-secreting gland or
78 M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora
elaiophore. Most Calceolaria species are pollinated
by oil-collecting bees (Vogel 1974, Se
´rsic 1994).
By contrast, in C. uniflora the upper edge of
the instep is folded outward as a transversal white
and fleshy appendage (Fig. 1), while the elaio-
phore is weakly developed and non-functional
(Se
´rsic and Cocucci 1996, Roitman et al. 2002).
This structure appears only in one additional,
closely related species, C. fothergillii Ait. of the
Falkland Islands and is the main diagnostic
character of the section Kremastocheilos (Witasek
1905, Ehrhart 2000). The appendage of C. uniflora
is fleshy and sweet (2% sugar in expressed juice)
and is consumed by birds (Thinocorus rumicivorus)
that in the process pollinate the flower (Se
´rsic and
Cocucci 1996). The species is largely allogamous,
though a few fruits and seeds can be produced by
autogamy (Arroyo and Squeo 1990, Masco
´et al.
2000, Masco
´2003).
Population sampling. Populations of C. uniflo-
ra throughout its range in Argentina were sampled
in the flowering season in December 1998, 1999 and
2000. Extensive exploration trips, covering over
4000 km of roads and tracks in three years, were
directed to areas where populations of the species
were known or suspected to exist. In total, 26
populations were sampled: 4 in 1998, 11 in 1999
and 11 in 2000. One population was repeatedly
sampled in both 1999 and 2000. The geographic
coordinates of each population were measured
using a Garmin 75 GPS. Topographic and vegeta-
tion data were recorded at each site. Normal mean
precipitation and temperature data were interpo-
lated from climatic maps (Conti 1998, Oliva et al.
2001) and meteorological station records (De Fina
et al. 1968). Monthly precipitation data for 1998 to
2000 were obtained from Mr. J. Aldridge, Rı
´o
Gallegos.
In every population, between 20 (in 1998) and
30–35 (in 1999 and 2000) flowers from different
plants were sampled at intervals of at least 5 m. The
total sample included 777 flowers for which com-
plete data were obtained.
Flower color photography. Each flower was
photographed frontally from a distance of 6.5 cm
using a Nikon FM10 camera, with a Nikon AF
Nikkor 60 mm lens. A Starblitz 1000 Auto Macro-
Lite annular flash was used to ensure standard and
uniform illumination.
The colors of the two frontal parts of the
flower were estimated separately:
– the instep, the external surface of the inflated
distal part of the lower corolla lip.
throat
appendage
instep
yellowred yellowred
0
1
2
3
cm
a
b
Fig. 1. Photographs of two flowers
of Calceolaria uniflora,showingthe
parts of the lower corolla lip, and
luminosity histograms of the instep
of each flower. aFlower with
uniform instep. bFlower with
maculate instep
M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora 79
– the throat, the internal surface of the lower lip,
visible above the instep (Fig. 1).
Objective color measurements on photo images.
An original objective quantitative method was
developed to estimate the proportions of surface
covered by different colors in the flower parts of
interest (Fig. 1). The photograph of each flower was
scanned at a resolution of 150 ppi and standard
settings and the image was analyzed using Corel
Photo Paint 8. The area of a flower part (instep or
throat) in the image was marked by a polygon. The
Corel Histogram function for RGB total luminosity
(on a scale from 0 ¼black to 255 ¼white) was
applied to the marked area, to read the proportion
of area (pixels) in three ranges of luminosity: <80
(dark), 80–120 (intermediate) and >120 (light)
(Fig. 1). In flowers of C. uniflora, these ranges
correspond approximately to the human-perceived
colors ‘‘red’’ (including red-brown), ‘‘orange’’ and
‘‘yellow’’, respectively, as was established by
marking small areas of uniform hue in the flower
image and reading their luminosity.
The yellow color of Calceolaria flowers is
associated with carotenoid pigments in the chro-
moplasts (Wrischer and Ljubesic 1984). The darker
red and brown patches indicate anthocyanins in the
vacuoles.The luminosity values we measured are
probably negatively correlated with anthocyanin
concentrations in a patch.
Statistical analysis. Discriminant analysis was
used to evaluate the coincidence between visually
identified types of color patterns and objectively
measured color variables. Analysis of variance
within and between populations was performed
on the arcsine transformed area proportions of red,
orange and yellow in the instep and in the throat.
The stability (vs. plasticity) of these color variables
was tested by the significance of differences between
samples obtained in different years, over all pop-
ulations and within a single population. Where
non-linear relationships were indicated, e.g. be-
tween color variables and geographic and climatic
variables, the Spearman rank correlation coefficient
was used to test for monotonic trends. Multivariate
analysis of populations by cluster analysis (Ward’s
minimum variance method, Di Rienzo et al. 2000)
and principal component analysis was performed
with a set of 7 color variables: the mean propor-
tions of the three colors in the instep and in the
throat, and the proportion of flowers with uniform
instep. Phenetic Euclidean distances between all
pairs of populations were calculated from the 6
color variables (proportions of red, orange and
yellow in instep and throat) and compared with
geographic distances by various tests. All statistical
analysis and most graphs were performed with
INFOSTAT software (Di Rienzo et al. 2000).
Results
Visual identification of color patterns. Even
before the quantitative analysis, we identified
two visually distinct types of color pattern in
the instep area:
uniform:the surface of the instep is uniform
red, in hues that range from red-orange to
wine red, red-brown and dark brown
(Fig. 1a);
maculate: the surface of the instep shows a
pattern of darker patches of red or red-
brown and lighter patches of yellow grading
to orange (Fig. 1b).
The proportion of the surface covered by
red and yellow patches in maculate flowers is
highly variable. The darker patches are in the
form of irregular spots (maculae) on the yellow
background, when they cover less than half of
the surface. When they cover more, they merge
into a reticulate pattern, forming a labyrinth of
contrasting red-brown and yellow corridors.
The two types of instep color pattern were
easily identifiable in the field and in photo-
graphs. Image analysis differentiated clearly
between the two types (Fig. 1). In flowers with
uniform instep, the histogram of luminosity
showed a single sharp peak with the mean in
the darker (<80) range, with a standard
deviation of 5 to 10 units. In flowers with
maculate instep, the histogram was bimodal or
strongly skewed, with a standard deviation of
15 to 50 units. In a discriminant analysis based
on objectively measured proportions of three
colors in the instep over all flowers (N ¼777)
in relation to the visual ‘‘instep type’’ classifi-
cation, there was 87% coincidence between
visual and discriminant classifications. The
histogram of positions of individual flowers
along the axis of the discriminant function
80 M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora
showed a strongly bimodal distribution
(Fig. 2).
In contrast to the instep, the inner surface
of the throat was always maculate, with
small spots of dark orange, red or brown
color aligned on vertical lines on a yellow to
light orange background (Fig. 1). The pro-
portions and hues of the darker and lighter
patches were variable, but no discrete types
of color pattern could be distinguished in the
throat.
In the total sample there were 217 (28%)
flowers with uniform and 560 (72%) with
maculate instep. The frequency of flowers with
uniform vs. maculate instep varied greatly
between populations (Fig. 3). In two popula-
tions all flowers were uniform and in three
additional populations more than 85% of
flowers were uniform. These 5 populations
were labeled uniform populations. In 11 pop-
ulations all flowers were maculate and in 5
additional populations more than 90% of
Discriminant function from maculate to uniform instep type
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0.11
Relative fre q uency
Fig. 2. Frequency distribution of
positions of individual flowers
(N ¼777) along discriminant
axis with instep color pattern
(uniform/maculate) as class and
three objective instep colors as
variables
0
10
20
30
40
50
60
70
80
90
100
uniform-mixed-maculate populations
percentage of flowers with instep type
% uniform % maculate
Fig. 3. Distribution of the fre-
quencies of uniform and macu-
late instep flowers over
C. uniflora populations, ranked
in decreasing frequency of uni-
form flowers
M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora 81
flowers were maculate; these were labeled
maculate populations. There were 5 mixed
populations where plants with flowers of both
instep types occurred in frequencies between
30 and 70%.
Character stability. The objectively mea-
sured proportions of red, orange and yellow in
the instep were not significantly different
(P >0:15) between populations sampled in
years with drought periods in spring (1998–
99) and in the year 2000, when spring precip-
itation was evenly distributed and flowers were
abundant. In the Rı
´o Rubens population,
sampled in both 1999 and 2000, the frequency
of uniform flowers remained constant (61%
and 58% respectively, P ¼0:80 with Fisher’s
test for independence, N ¼62) and there were
no significant differences between years in the
proportions of the three instep colors.
As to throat colors, populations sampled in
2000 showed a slightly lower proportion of
orange and greater proportion of yellow than
those sampled in 1998–99. The difference was
about 7% and was significant (P ¼0:0182)
only for orange. A similar minor shift was
observed in the Rı
´o Rubens population
between 1999 and 2000 (P ¼0:0379).
Variance within and between popula-
tions. The between-populations component
accounted for over 50% of total variance in
the proportions of red and yellow in the instep;
the mean over the three instep colors was 48%
(Table 1). Throat colors had relatively more
within-population and less between-popula-
tions variance (mean 24%).
Correlation between instep and throat
colors. The proportion of different colors in
the throat did not differ significantly between
instep color types (N ¼777 flowers). All
Spearman correlation coefficients between the
proportions of the three main instep colors and
the three main throat colors were not signif-
icant over all populations (P >0:10, N ¼26),
indicating overall independent color variation
in the two flower parts (Fig. 4).
However, within the set of maculate pop-
ulations (N ¼16) there was a tendency for
coincident variation in color proportions in the
throat and the instep (Fig. 4). Yellow in the
throat was positively correlated (rho ¼þ0:61)
with yellow in the instep, and orange in the
throat was positively correlated (þ0.66) with
red in the instep (P <0:05).
Geographic trends in flower colors. Instep
color pattern types showed a strong longitudi-
nal pattern (Fig. 5): populations with uniform
instep occurred at or west of 72°W, mixed
populations occurred between 71 and 72°W,
and only maculate populations were found
east of 71°W. The frequency of flowers with
uniform instep was strongly but not linearly
correlated with longitude west (Spearman
Table 1. Components of variance between and within 26 populations in the objectively measured area
proportions of colors in the instep and throat of Calceolaria uniflora flowers (N ¼777). Angular trans-
formation was applied to the proportions. The between population variance was highly significant
(P <0:0001) in all cases
Color variable Luminosity range % variance between % variance within
Instep
Red <¼80 52.2 47.8
Orange 80–120 34.8 65.2
Yellow >120 56.8 43.2
Mean 47.9 52.1
Throat
Red <¼80 21.0 79.0
Orange 80–120 26.9 73.1
Yellow >120 25.0 75.0
Mean 24.3 75.7
82 M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora
rho ¼þ0:78, P <0:0001, Table 2, Fig. 6).
This longitudinal limit between uniform and
maculate instep populations corresponded to
the biogeographic limit between Andean-su-
bandean forest/grassland formations and
Magellanic steppe formations (Oliva et al.
2001). Mixed populations occurred in the
southern Andean-subandean region, where
Subandean grassland grades into Magellanic
steppe. The frequency of flowers with uniform
instep was not significantly correlated with
latitude and only weakly with altitude (Ta-
ble 2).
The proportions of red, orange and yellow
in the instep were strongly (jrho0:65 to
0.70) and significantly (P <0:01, P <0:001)
correlated with longitude (Table 2). The trend
was stepwise rather than linear: between 71
and 72°W the proportion of red increased
sharply towards the west, while the propor-
tions of yellow and orange increased towards
the east. Instep colors did not show any trend
with either latitude or altitude over the entire
sample (Table 2). However, within the sub-
sample of maculate populations (N ¼16), red
in the instep increased and yellow decreased
significantly with latitude south (jrho0:66,
P<0:05).
Throat colors showed no significant longi-
tudinal trend but did vary with latitude. The
proportion of orange in the throat increased
significantly from north to south, while yellow
decreased in the same direction (Table 2). This
latitudinal trend in throat colors was mani-
fested mainly within maculate populations.
Correlation of flower colors with climatic
variables. The proportion of uniform instep
flowers in the population was not significantly
correlated with temperature (P >0:20) and
Maculate Mixed Uniform instepMaculate Mixed Uniform instep
010 20 30 40 50 60
%yellowininstep
30
40
50
60
70
80
90
% ye llo w i n throa t
Fig. 4. Mean % of yellow in throat vs. mean % of
yellow in instep in populations of C. uniflora
(N ¼26) by instep color type: black triangles –
uniform, light gray circles – maculate, dark gray
squares – mixed populations
Río Gallegos
Pta. Arenas
TIERRA DEL FUEGO
Estrecho de Magallanes
SANTA CRUZ
CHILE
R.SantaCruz
Andes
72ºW
52ºS
68ºW
48ºS
0 200 km
an
d
ina
magellanica
Fig. 5. Map of study area and populations of
C. uniflora by instep type: black triangles – uniform,
light gray circles – maculate, dark gray squares –
mixed populations. Large symbols – populations
sampled, small symbols – other populations. Dotted
lines: northern and eastern limit of C. uniflora,and
approximate limits of uniform, maculate and mixed
populations
M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora 83
only weakly correlated with precipitation
(rho ¼þ0:46, P ¼0:0222). However, there
were significant correlations between tempera-
ture and increased proportions of yellow and
orange in the instep and of yellow in the throat
(Table 3). Conversely, the proportions of red in
the instep and of orange in the throat increased
in colder sites. Increased precipitation was
significantly correlated with increasing red in
instep and orange in throat and with decreasing
yellow in instep.
In general, color variables were not as
strongly correlated with climatic variables
(jrho0:50:6) as they were with longitude
and latitude (jrhoj>0:6; Table 2). Though
only maculate populations occurred in the
driest and warmest parts of the species range,
there was a range of temperature (10–11 °C)
and precipitation (250–350 mm) where uni-
form, maculate and mixed populations over-
lapped.
Multivariate analysis. Cluster analysis
(with 7 variables: 3 instep, 3 throat colors, %
of uniform instep) separated at the highest
level all Andean-subandean populations of
uniform or mixed instep type (N ¼10) from
all Magellanic populations with maculate
instep (N ¼16) (Fig. 7). The second division
of the first group segregated uniform from
mixed populations. The second division of the
maculate group separated populations with
lighter colors in both throat and instep (mostly
in the north) from populations with darker
colors (mostly in Tierra del Fuego and in the
south). One population with a very high
proportion of yellow was set apart as an
outlier group and was excluded from the
principal components analysis.
In PCA of the 25 populations using the
same 7 variables, the first principal component
Table 2. Spearman rank correlation coefficients (rho) between site geographic variables (latitude, longitude
and altitude) and site (population) means of color variables in the instep and throat of Calceolaria uniflora
flowers (N ¼26 sites). Significance levels: NS – P >0:05; –P<0:05;  –P<0:01;  –P<0:001.
Significant (P <0:05) values highlighted in bold type
Latitude Longitude Altitude
Instep type
% of uniform flowers 0.30 NS 0.78 *** 0.48 *
Instep colors
Red (<80) 0.05 NS 0.65 ** 0.21 NS
Orange (80–120) 0.18 NS 0.70 *** 0.38 NS
Yellow (>120) 0.16 NS 0.67 *** 0.27 NS
Throat colors
Red (<80) 0.16 NS 0.18 NS 0.00 NS
Orange (80–120) 0.57 ** 0.25 NS 0.44 *
Yellow (>120) 0.44 * 0.17 NS 0.29 NS
Maculate Mixed Uniform instepMaculate Mixed Uniform instep
73 72 71 70 69 68 67
Longitude W
0
20
40
60
80
100
% of flowers with uniform instep
Fig. 6. Percentage of uniform-instep flowers in
C. uniflora populations vs. longitude West (scale
inverted to fit map vision), by instep color type.
Symbols as in Fig. 4
84 M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora
accounted for 51% of variance and the second
one for 36%. The first axis expressed mainly
variation in instep color, contrasting uniform
instep and high proportion of red at the left
extreme with maculate instep and high
proportions of yellow and orange at the right
extreme (Fig. 8). The second axis expressed
mainly variation in throat color, with more of
red and orange at the lower extreme and more
of yellow at the upper extreme. The three
groups of populations defined by instep type,
uniform, maculate and mixed, were clearly
clustered in ordination space.
The first axis showed a strong correlation
with longitude (rho ¼0:70, P ¼0:0006)
and a weaker correlation with temperature
(rho ¼þ0:51, P ¼0:0126). The populations
clustered on the left end of the first axis (instep
colors) were mostly in Andean-subandean
vegetation, while those at the right end were
in Magellanic steppe. The second axis (throat
colors) showed a weak trend with latitude
(rho ¼0:44, P ¼0:0302), and a stronger
correlation with temperature (rho ¼þ0:61,
P¼0:0027).
Geographic distance and phenetic dis-
tance. Only 17% of population pairs within
less than 50 km distance were of different
instep color type, compared to 43% of pairs
at distances of 50 to 150 km, and 61% of
pairs that were more than 150 km distant
(Chi-square for independence ¼16:3, df ¼2,
P¼0:0003, N ¼276). The correlation be-
tween the phenetic distance calculated from 6
color variables and the geographic distance
was positive (rho ¼þ0:405) and highly
significant (P ¼0:0010; distances <150 km;
Table 3. Spearman rank correlation coefficients (rho) between normal climatic variables (mean January
and mean annual temperature, annual precipitation) and site means of area proportions of colors in the
instep and throat of Calceolaria uniflora flowers (N ¼26 sites). Significance levels: NS P >0:05;
*–P<0:05; ** – P <0:01; *** P <0:001. Significant (P <0:05) values highlighted in bold type
Mean January
temperature
Mean annual
temperature
Annual
precipitation
Instep colors
Red (<80) 0.50 *0.51 *0.58 **
Orange (80–120) 0.49 *0.63 ** 0.36 NS
Yellow (>120) 0.57 ** 0.58 ** 0.61 **
Throat colors
Red (<80) 0.24 NS 0.36 NS 0.01 NS
Orange (80–120) 0.57 ** 0.51 *0.46 *
Yellow (>120) 0.57 ** 0.61 ** 0.20 NS
0 5 10 15 20 25
Phenetic distance
BelVista
MarcachA
Violetas
PSina
i
CalMision
SSebas
t
R9
RPTero
Esc
GuerAcur
LeMarch
MLeon
GuerAike
MDinero
LagAzul
MoyAike
PMEnt
PNGlacia
PMLAOr
PMLBur
VaArg
RioRub
SantAna
RupP
N
RupPS
Zurdo
Fig. 7. Dendrogram of cluster analysis of C. uniflora
populations (N ¼26) by Ward’s minimum variance
method with seven color variables. Bold type –
uniform instep populations, bold italic – mixed
populations, regular type – maculate populations on
the continent, italic – maculate populations on Tierra
del Fuego
M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora 85
N¼76). The mean phenetic distance between
populations within 25 km of each other was
only 0.17, compared to >0.50 between popu-
lations at 100 to 150 km distance (Fig. 9;
ANOVA: P ¼0:0324, R2¼0:10, N ¼76).
Discussion
Patterns of variation in flower colors. Differ-
ences in morphological characters between
geographically distant populations are ideally
evaluated in common garden experiments.
However, this is not always feasible with
perennial species that are difficult to grow
and maintain in culture. Many biosystematic
studies are based on in situ measurements in
wild populations (e.g. Domı
´nguez et al. 1998)
or on herbarium specimens (e.g. Lefebvre and
Vekemans 1995). This creates a problem in
interpreting observed differences between pop-
ulations. To what extent do the latter reflect
genetic differences and to what extent – plastic
phenotypic responses to environmental condi-
tions at the site and time the population
was sampled? Differences between in situ
populations are reliable indicators of genetic
differences only for characters that are pheno-
typically stable under normal variations in
environmental conditions. In this study, com-
parison between samples taken in years with
different precipitation indicated that flower
BelVista
CalMision
Esc
GuerAcu
GuerAike
LagAzul
LeMarch
MarcachA
MDinero
MLeon
PMEnt
PMLAOr
PMLBur
PNGlacia
PSinai
R9
RioRub
RPTero
RupPN
RupPS
SantAna
SSebast
VaAr
g
Violetas
Zurdo
BelVista
CalMision
Esc
GuerAcu
GuerAike
LagAzul
LeMarch
MarcachA
MDinero
MLeon
PMEnt
PMLAOr
PMLBur
PNGlacia
PSinai
R9
RioRub
RPTero
RupPN
RupPS
SantAna
SSebast
VaAr
g
Violetas
Zurdo
%Uniform Ins tep
%Red Instep
%Orange Instep
%Yellow Instep
%Red Throat
%Orange Throat
%Yellow Throat
%Uniform Ins tep
%Red Instep
%Orange Instep
%Yellow Instep
%Red Throat
%Orange Throat
%Yellow Throat
PC 1
PC 2
Fig. 8. Biplot ordination of
color variables (vectors with
black triangles) and popula-
tions of C. uniflora (circles;
N¼25) by first (PC1) and
second (PC2) principal com-
ponents from correlations be-
tween seven color variables.
Black circles – uniform instep,
dark gray – mixed, light gray
– maculate populations
<25 25-50 50-7 5 75-100 100-125125-150
Geographic distance (km)
0.0
0.1
0.2
0.4
0.5
0.6
Phenetic distrance (6 color variables)
Fig. 9. Mean and standard error of phenetic dis-
tances, calculated as Euclidean distances from six
color variables, between C. uniflora populations vs.
classes of geographic distances between them, up to
150 km distance
86 M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora
colors and color patterns in C. uniflora are
relatively stable. In contrast, some of the
quantitative morphometric variables, such as
flower dimensions, showed significant plastic
responses to year and site conditions (Masco
´
2003). Evidence from other species suggests
that the presence of a flower color pattern (e.g.
spots) is often phenotypically stable and is
determined by a single major gene (e.g. Jones
1996, Mol et al. 1998). However, the expres-
sion of color and its intensity can be influenced
by current environmental conditions (e.g.
Weiss 2000).
C. uniflora is an outcrossing species (Masco
´
2003), in which high variation between indi-
viduals within local populations can be ex-
pected. In the outcrossing Ipomopsis
aggregata, 70–80% of the variance in different
flower morphometric variables was within
populations and only 20–30% between popu-
lations (Wolf and Campbell 1995). In
C. uniflora flowers, between-populations vari-
ance in the proportions of red and yellow in
the instep was relatively high (>50%, Table 1).
The stable and large between-populations
variation suggests that instep colors and color
patterns are diagnostic characters for geo-
graphic variation within C. uniflora.
Throat colors showed lower between-pop-
ulations variance (20–25%). Over all popula-
tions, throat colors varied independently of
instep colors (Figs. 4, 8), indicating that dif-
ferent genes may be involved. In flowers with
radial symmetry, throat and limb color often
vary independently (e.g. Clegg and Durbin
2000, Wolfe 2001).
Infraspecific taxa? To what extent is var-
iation within a species either discontinuous or
continuous in character space and in geo-
graphic space (Prentice 1986)? Where it is
discontinuous or strongly clustered, there may
be justification for distinguishing infraspecific
taxa. Where it is largely continuous, any
subclassification will be arbitrary and it is
more meaningful to define clines of variation.
Within C. uniflora, uniform and maculate
instep populations formed distinct clusters in
character space (Fig. 8) and were clearly
segregated in geographic space (Fig. 5). Addi-
tional qualitative information on instep color
pattern from another 14 sites in Argentina and
Chile and from herbarium specimens con-
firmed this pattern. No overlap between the
geographic distributions of pure uniform and
maculate populations was found, though
mixed populations occurred in an intermediate
zone. Uniform populations differed signifi-
cantly from maculate populations in four
phenotypically stable morphometric characters
of flower size and shape (Masco
´2003).
According to the criteria of Du Rietz (1930;
cited by Stace 1989), ‘‘subspecies’’ are mor-
phological races differentiated at a regional
scale. Thus, there is preliminary evidence for
defining two ‘‘subspecies’’ within C. uniflora:1)
populations with uniform red corolla instep,
distributed along the Andes (informally la-
beled ‘‘andina’’); 2) populations with maculate
yellow-red instep, distributed in the steppe on
both sides of the Strait of Magellan (‘‘magel-
lanica’’). Between the core areas of the two
subspecies there is a phenocline of mixed
populations. The phenocline could indicate a
focus of recent divergence of the two subspe-
cies, or it may be the result of a secondary
contact and hybridization between taxa that
diverged long ago.
A closely related taxon is C. fothergillii
Ait., that shares with C. uniflora the white
transversal appendage on the lower corolla lip.
This species is reliably documented only from
the Falkland Islands, where it is isolated by
500 km of ocean from the nearest populations
of C. uniflora on the continent. The color
pattern on the instep of C. fothergillii, as seen
in flower images (Curtis 1796, Vallentin and
Cotton 1921, Davies and McAdam 1989,
Woods 2000) is quite distinct from that of
both subspecies of C. uniflora.
Mechanisms of differentiation: pollinators,
climate or distance? What evolutionary mech-
anisms generated and maintained the geo-
graphic differentiation in floral color patterns
within C. uniflora? A prevalent hypothesis
(since Darwin 1859) is that selection by
different pollinators in different regions can
M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora 87
trigger and stabilize ‘‘adaptive’’ differentiation
in flower colors. Experimental evidence sup-
porting this mechanism has been frequently
reported (e.g. Vickery 1992, Schemske and
Bradshaw 1999, Medel et al., in press). The
only known pollinator of C. uniflora is the
seedsnipe Thinocorus rumicivorus (Se
´rsic and
Cocucci 1996) that consumes the appendage
in a large proportion of flowers and appar-
ently induces high fruit set observed in
natural populations (Masco
´2003). The large
(1,1 X 0,5 cm) and brilliant white append-
age is the reward as well as a specific visual
signal for the pollinator. The red, orange and
yellow surfaces of the lower lip produce with
the white band a striking color combination,
that in the open Patagonian landscape is
highly visible to the human eye, as it probably
is to birds. The high proportion of red
(anthocyanin) surfaces in the flowers of
C. uniflora, compared to sympatric species
(most Calceolaria species are yellow with at
most small reddish spots) could have evolved
as an ‘‘advertisement’’ that increased the
visibility of the flower to birds foraging on
the ground for seeds and berries. But there is
no evidence to attribute to pollinator selection
the differentiation between the uniform red
instep Andean populations and the maculate
yellow-red Magellanic populations. Thinocor-
us rumicivorus is a migratory bird that in
spring and summer is common throughout
the Magellanic steppe. It is less common in
the Andean region, but there it is replaced by
other seedsnipes (Thinocorus orbignyanus, At-
tagis gayi, A. malouinus; A. Manero, pers.
comm.). Flowers with damaged appendage
were observed also in some Andean popula-
tions (Se
´rsic and Cocucci 1996 and personal
observations). The Andean seedsnipes proba-
bly function as pollinators in much the same
way as T. rumicivorus in the steppe. It seems
unlikely that closely related bird taxa select
for different color patterns in the two regions.
The hypothesis that a larger amount of yellow
surfaces in Magellanic populations is main-
tained through selection by alternative polli-
nators (pollen-collecting bees?) cannot be
rejected but is speculative in the absence of
supporting evidence.
An alternative ‘‘adaptive’’ hypothesis is
that geographic differentiation in C. uniflora
flower colors resulted from selection by factors
other than pollinators, in particular climate.
The presence and concentration of anthocya-
nins in leaves and stems are often correlated
with drought, cold or UV-radiation stress,
leading to the hypothesis that the primary
adaptive role of these pigments may have been
protection from stress (Warren and Mackenzie
2001). In some taxa, their concentration in
flowers may have resulted from indirect selec-
tion by climatic stress rather than selection by
pollinators (Armbruster 2002). In this study we
found a significant correlation between in-
crease in red (anthocyanin) areas in both parts
of the corolla and colder temperatures in the
south (Table 3). UV radiation probably in-
creases along the same gradient. The increase
in anthocyanin may protect flowers and leaves
from cold and radiation damage.
Another possible mechanism for this lati-
tudinal trend could be introgression from the
yellow-flowered Calceolaria polyrhiza Cav.
This species, although from a different section
than C. uniflora and lacking the white append-
age, hybridizes with it naturally throughout
the northern Magellanic region (Se
´rsic et al.
2001) but was not found south of the
R. Gallegos nor in Tierra del Fuego.
The longitudinal differentiation in instep
type between Andean and Magellanic popula-
tions is not easily explained by climatic or
pollinator factors alone. The proportion of
flowers with uniform instep was not signifi-
cantly correlated with temperature and only
weakly with precipitation. The climatic ranges
of populations with different instep types
overlapped much more than their geographic
ranges. The distribution of instep types in the
multivariate space defined by color variables
was remarkably similar to the geographic
distribution (Figs. 8 vs. 5). A weak but signif-
icant correlation was found between phenetic
and geographic distances between populations
up to 150 km (Fig. 9). The color characters of
88 M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora
some populations seem to reflect geographic
proximity and continuity rather than site
conditions. For instance, the northeastern
coast of Tierra del Fuego is as cold and rainy
as some of the Andean sites but harbors
maculate populations very similar to those in
the continental steppe (Figs. 5, 7). At the
northwestern extreme of a large steppe region
where ‘‘magellanica’’ populations are nearly
continuously distributed, a typically maculate
population inhabits one of the highest and
coldest sites in our sample (Los Escarchados).
It resembles ‘‘magellanica’’ populations to the
east and south rather than ‘‘andina’’ popula-
tions to the west, across the dry Santa Cruz
Valley where C. uniflora is absent. On the other
hand, uniform ‘‘andina’’ populations with little
color variation inhabit various microclimates
along the Andes, from high mountain semi-
desert to Subandean grassland and clearings in
lowland Nothofagus forest.
These observations suggest that gene flow,
isolation by distance and by geographic barri-
ers have had a role in generating the present
geographic variation in flower colors in
C. uniflora. The original divergence between
‘‘andina’’ and ‘‘magellanica’’, whether ancient
or recent, may have involved selection by the
colder and wetter Andean climate vs. the drier
climate of the Magellanic steppe, or by different
pollinators. These factors may still be operating
in the phenocline, where mixed populations can
serve as stepping-stones for gene flow between
uniform and maculate populations. However,
where geographic distance and discontinuities
in distribution restrict gene flow, populations
tend to be most similar to the nearest popula-
tions from which they originated and with
which they maintain gene flow.
There is no direct information on gene flow
in this species. Movements of the pollinating
birds (Thinocorus spp.) are mostly local in the
flowering season, but the seedsnipe is capable
of occasionally transporting pollen between
populations over distances of a few kilometers.
Another mechanism of gene flow over medium
distances can be the dispersal of the tiny seeds
(0.05 mg; Masco
´2003) of C. uniflora by the
extremely strong winds, mainly from the west,
that characterize the Southern Patagonian
summer. The similarity between Fuegian and
continental populations on opposite sides of
the Strait of Magellan suggests that the latter is
not an effective barrier to gene flow. The Strait
is about 9,000 years old and has an average
width of about 30 km but narrows to 4–7 km
at two points. By contrast, an arid salient
about 80 km wide seems to effectively isolate
the northernmost maculate populations from
uniform Andean populations. These two
observations indicate a scale for gene flow in
this species.
This was part of the research for a M.Sc. thesis
of the first author at the Faculty of Agricultural
Sciences, National University of Co
´rdoba, Argen-
tina, supported in part by a FOMEC scholarship.
The third author was supported by CONICET.
Fieldwork was supported by the EEA (INTA)
Santa Cruz and a grant from GTZ (Germany). We
are grateful to Liliana Gonza
´lez, Eduardo Quarg-
nolo, Gabriel Oliva and Daniel Barrı
´a who helped
in various ways to realize this research.
References
Armbruster W. S. (2002) Can indirect selection and
genetic context contribute to trait diversifica-
tion? A transition-probability study of blossom-
colour evolution in two genera. J. Evol. Biol. 15:
468–486.
Arroyo M. T. K., Squeo F. (1990) Relationship
between plant breeding systems and pollination.
In: Kawano S. (ed.) Biological approaches and
evolutionary trends in plants. Academic Press,
London, pp. 205–227.
Boelcke O., Correa M. N., Moore D., Roig F.
(1985) Cata
´logo de plantas Vasculares. In: Boel-
cke O., Moore D., Roig F. (eds.) Transecta
bota
´nica de la Patagonia Austral. CONICET,
Instituto de la Patagonia & Royal Society,
Buenos Aires, pp. 129–255.
Chittka L., Shmida A., Troje N., Menzel R. (1994)
Ultraviolet as a component of flower reflections,
and the color perception of Hymenoptera.
Vision Res. 34: 1489–1508.
M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora 89
Clegg M. T., Durbin M. L. (2000) Flower color
variation: A model for the experimental study of
evolution. Proc. Natl. Acad. Sci. USA 97: 7016–
7023.
Conti H. A. (1998) Caracterı
´sticas clima
´ticas de la
Patagonia. In: Correa M. N. (ed.) Flora Patago
´-
nica, Pt. I. Coleccio
´n Cientı
´fica, INTA, Buenos
Aires, pp. 31–47.
Correa, M. N. (1999) Flora Patago
´nica.Pt. VI.
Dicotyledones Gamopetalas (Ericaceae a Caly-
ceraceae). Coleccio
´n Cientı
´fica, INTA, Buenos
Aires, pp. 1–536.
Curtis W. (1796) Curtis’ flower prints. Bot. Mag.,
London 10: 348.
Darwin C. (1859) The origin of species. Reprint of
Oxford University Press, Oxford, pp. 1–437.
Davies T. H., Mc Adam J. H. (1989) Wild flowers
of the Falkland Islands. Falkland Islands Trust,
Stanley, Falkland Islands, pp. 1–48.
Descole H., Borsini O. (1954) Scrophulariaceae. In:
Descole H. (ed.) Genera et species plantarum
Argentinarum. Guillermo Kraft, Buenos Aires,
Vol. 5., pp. 1–92.
De Fina A., Garbosky A., Giaunetto F., Sabella L.
(1968) Difusio
´n geogra
´fica de los cultivos ı
´ndices
en la provincia de Santa Cruz y sus causas.
Publicaciones Te
´cnicas, INTA, Buenos Aires,
pp. 1–30.
Di Rienzo J., Robledo W., Casanoves F., Balza-
rini M. (2000) InfoStat, Software estadı
´stico.
Facultad de Ciencias Agropecuarias, Universi-
dad Nacional de Co
´rdoba, Co
´rdoba,
Argentina.
Domı
´nguez C. A., Eguiarte L. E., Nu´ n
˜ez-Farfa
´n J.,
Dirzo R. (1998) Flower morphometry of Rhizo-
phora mangle (Rhizophoraceae): geographic var-
iation in Mexican populations. Amer J. Bot. 85:
637–643.
Du Rietz G. E. (1930) The fundamental units of
biological taxonomy. Svensk. Bot. Tidskr. 24:
333–428.
Ehrhart C. (2000) Die Gattung Calceolaria
(Scrophulariaceae) in Chile. Bibliotheca Botan-
ica 153: 1–283.
Elkington T. T. (1986) Patterns of variation in
wild plants in relation to selection pressures
and infraspecific categories. In: Styles B. T.
(ed.) Infraspecific classification of wild and
cultivated plants. The Systematics Association,
Special Vol. 29: 33–52. Clarendon Press,
Oxford.
Faegri K., van der Pijl I. (1979) The principles of
pollination ecology, 3
rd
edn. Pergamon, Oxford,
pp. 1–244.
Jones K. N. (1996) Fertility selection on a discrete
floral polymorphism in Clarkia (Onagraceae).
Evolution 50: 71–79.
Judd W., Campbell C., Kellogg E., Stevens P.
(1999) Plant systematics. A phylogenetic ap-
proach. Sinauer Associates, Inc. Sunderland,
pp. 1–464.
Lefebvre C., Vekemans X. (1995) A numerical
taxonomic study of Armeria maritima (Plumba-
ginaceae) in North America and Greenland.
Canad. J. Bot. 73: 1583–1595.
Masco
´M. (2003) Sistema reproductivo y variabil-
idad floral infraespecı
´fica en Calceolaria uniflora
Lam., una especie ornamental ende
´mica de la
Patagonia austral. M. Sc. Thesis. Universidad
Nacional de Co
´rdoba, Argentina, pp. 1–108.
Masco
´M., Oliva G., Kofalt R., Humano G. (1998)
Flores Nativas de la Patagonia Austral. Una
seleccio
´n de especies silvestres con potencial
ornamental de Santa Cruz y Tierra del Fuego.
Montes, L. (ed). EEA-INTA Santa Cruz,
pp. 1–48.
Masco
´M., Se
´rsic A. N., Noy-Meir I. (2000)
Observaciones sobre la biologı
´a reproductiva y
diversidad infra-especı
´fica de Calceolaria uniflora
(Scrophulariaceae). Gayana Botanica 57: 93.
Medel R., Botto-Mahan C., Arroyo M. T. K.
(2003) Pollinator-mediated selection on the nec-
tar guide phenotype in the Andean monkey
flower, Mimulus luteus. Ecology (in press).
Mele
´ndez-Ackerman E. J. (1997) Pattern of color
and nectar variation across an Ipomopsis (Pole-
moniaceae) hybrid zone. Amer. J. Bot. 84: 41–47.
Moore D. (1983) Flora of Tierra del Fuego.
Anthony Nelson, London, pp. 1–369.
Mol J., Grotewold E., Koes R. (1998) How genes
paint flowers and seeds. Trends in Plant Science
3: 212–217.
Molau U. (1988) Scrophulariaceae. Part I. Cal-
ceolarieae. Flora Neotropica. The New York
Botanical Garden, Bronx, New York, pp. 1–326.
Oliva G., Gonzalez L., Rial P., Livraghi E. (2001)
El ambiente en la Patagonia Austral. In: Borrelli
P., Oliva G. (eds.) Ganaderı
´a sustentable en la
Patagonia Austral. EEA Santa Cruz, INTA, Rı
´o
Gallegos, pp. 41–82.
Podolsky R. H., Holtsford T. P. (1995) Population
structure of morphological traits in Clarkia
90 M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora
dudleyana. I. Comparison of F
st
between allo-
zymes and morphological traits. Genetica 140:
733–744.
Prentice H. C. (1986) Continuous variation and
classification. In: Styles B. T. (ed.) Infraspecific
classification of wild and cultivated plants. The
Systematics Association, Special Vol. 29: 21–32.
Clarendon Press. Oxford.
Roig F. A. O. (1998) La vegetacio
´n de la Pata-
gonia. In: Correa M. N. (ed.) Flora Patago
´nica,
Pt. I. Coleccio
´n Cientı
´fica, INTA, Buenos Aires,
pp. 48–166.
Roitman G. G., Se
´rsic A. N., Cocucci A. A.,
Montaldo N. H. (2002) Flower food tissue as
reward for pollinating birds. Bol. Soc. Argent.
Bot. 37: 71–77.
Schemske D. W., Bierzychudek P. (2001) Perspec-
tive: evolution of flower color in the desert
annual Linanthus parryae: Wright revisited.
Evolution 55: 1269–1282.
Schemske D. W., Bradshaw H. D. Jr. (1999)
Pollinator preference and the evolution of floral
traits in monkeyflowers (Mimulus). Proc. Natl.
Acad. Sci. USA 96: 11910–11915.
Se
´rsic A. (1994) Estudios de la biologı
´a floral del
ge
´nero Calceolaria L. Ph. D. Thesis, Universidad
Nacional de Co
´rdoba, Argentina.
Se
´rsic A., Cocucci A. A. (1996) A remarkable case
of ornithophily in Calceolaria: food bodies as
rewards for a non-nectarivorous bird. Botanica
Acta 109: 172–176.
Se
´rsic A., Masco
´M., Noy-Meir I. (2001) Natural
hybridization between species of Calceolaria
with different pollination syndromes in southern
Patagonia, Argentina.Plant Syst. Evol. 230:
111–124
Solbrig O. (1970) Principles and methods of plant
biosystematics. MacMillan, New York, pp.
1–226.
Stace C. A. (1989) Plant taxonomy and biosyste-
matics, 2
nd
edn. Edward Arnold, London, pp.
1–279.
Vallentin E. F., Cotton E. M. (1921) Illustrations
of the flowering plants and ferns of the Falkland
Islands. L. Reeve, London, pp. 1–130.
Vickery R. K. (1992) Pollinator preferences for
yellow, orange and red flowers of Mimulus
verbenaceus and M. cardinalis. Great Basin
Naturalist 52: 145–148.
Vogel S. (1974) O
¨lblumen und o
¨lsammelnde Bie-
nen. Abh. Akad. Wiss. Lit., Math. Naturwiss.
Kl. 7: 285–547.
Vorobyev M., Gumbert A., Kunze J., Giurfa M.,
Menzel R. (1997) Flowers through insect eyes.
Israel J. Plant Sci. 45: 93–101.
Warren J., Mackenzie S. (2001) Why are all colour
combinations not equally represented as flower-
colour polymorphisms? New Phytol. 151: 237–
241.
Weiss D. (2000) Regulation of flower pigmentation
and growth: Multiple signaling pathways control
anthocyanin synthesis in expanding petals. Phys-
iol. Pl. 10: 152–157.
Witasek J. (1905) Die chilenische Arten der Gattung
Calceolaria.O
¨st. Bot. Zeitschrift 55: 449–456.
Wolf P. G., Campbell D. R. (1995) Hierarchical
analysis of allozymic and morphometric varia-
tion in a montane herb, Ipomopsis aggregata
(Polemoniaceae). J. Heredity 86: 386–394.
Wolfe L. M. (2001) Associations among multiple
floral polymorphisms in Linum pubescens (Lin-
aceae), a heterostylous plant. Int. J. Plant Sci.
162: 335–342.
Woods R. (2000) Flowering plants of the Falkland
Islands. Falklands Conservation, Stanley, pp.
1–108.
Wrischer M., Ljubesic N. (1984) Plastid differenti-
ation in Calceolaria petals. Acta Bot. Croat. 43:
19–24.
Addresses of the authors: Ing. Agro
´n. Mercedes
Masco
´(e-mail: mermasco@correo.inta.gov.ar),
EEA Santa Cruz, Consejo Agrario Provincial,
INTA y Universidad Nacional de la Patagonia
Austral, C.C. 332, Rı
´o Gallegos 9400, Santa Cruz,
Argentina. Prof. Imanuel Noy-Meir (e-mail: noy-
meir@agri.huji.ac.il), Institute of Plant Sciences,
Faculty of Agricultural, Food and Environmental
Quality Sciences, Hebrew University of Jerusalem,
P.O.B. 12, Rehovot 76100, Israel. Dra. Alicia N.
Se
´rsic (e-mail: asersic@com.uncor.edu), Instituto
Multidisciplinario de Biologı
´a Vegetal (IMBIV),
Universidad Nacional de Co
´rdoba y CONICET,
C.C. 495, Co
´rdoba 5000, Argentina
M. Masco
´et al.: Geographic variation in flower colors within Calceolaria uniflora 91
... Additionally, variation in flower colour has commonly been interpreted as adaptive. The differentiation in flower colour therefore was considered an important factor in promoting plant speciation (Bradshaw et al. 1995;Matsumura et al. 2006;Hopkins and Rausher 2011) or promoting adaptive, resulting from the disruptive selection by different pollinators (Mascó et al. 2004;Irwin and Strauss 2005;Veiga et al. 2015). On the other hand, the reproductive system may influence the patterns of variation in some taxa, and might account for the morphological complexity. ...
Article
Full-text available
Morphological variations, particularly flower colour, could be considered as an evolutionarily and ornamentally significant taxonomic criterion for Epimedium . Our extensive field investigation based on population studies revealed abundant intraspecific variations in flower colour. Five species, (i.e., E.acuminatum Franch., E.leptorrhizum Stearn, E.pauciflorum K.C.Yen, E.mikinorii Stearn, and E.glandulosopilosum H.R.Liang) were found to possess polymorphic flower colour, which is first described and illustrated here. Moreover, all these species were found to be polymorphic in other diagnostic characters, such as the type of rhizome, the number and arrangement of stem-leaves, and/or their indumentum, which have not been adequately described in previous studies. Therefore, their morphological descriptions have been complemented and/or revised. We also provide notes on the morphology and nomenclature for each species. Additionally, a key to the species in China has been provided. The present study could serve as a basis for understanding their taxonomy and helping their utilisation as an ornamental plant.
... Few phenotypic traits, such as floral colour patterns and plant pilosity, differentiate these species (Sérsic and Cocucci 1996). From a reproductive perspective, the species display low fruit and seed set by autogamy, relying on bird pollination for pollen transfer (Arroyo and Squeo 1990;Sérsic 2004;Mascó et al. 2004). Because seeds are dispersed mainly by gravity and wind (Molau 1988;Ehrhart 2000;Fernández et al. 2002), gene flow among populations via seeds is presumably limited. ...
Article
Full-text available
A key to understanding the origin and identity of young species lays on the knowledge of the Quaternary climatic oscillations’ effect on gene flow and vicariance. Even though the effect of climatic fluctuations is relatively well understood for southern hemisphere plant species, little is known about their effect on the evolutionary histories of species from mainland and islands. Thus, we investigated whether Quaternary climate-driven fluctuations translated into lineage divergence and speciation, followed or not by climatic niche differentiation, in two allopatric plant species, Calceolaria uniflora and C. fothergillii from Patagonia and Malvinas/Falkland islands, respectively. We sampled the range of both species, and sequenced two chloroplastic (cpDNA; trnS–trnG and trnH–psbA), and one single copy “anonymous” non-coding nuclear region (nDNA). We performed phylogeographic and dating analyses, and adjusted spatio-temporal diffusion models. We complemented molecular evidence with climatic niche differentiation analyses and species paleo-distribution projections. A species coalescent reconstruction based on multi-locus data retrieved both species as monophyletic. Estimates from cpDNA indicated the species diverged during the Great Patagonian Glaciation. Chloroplast and nuclear DNA showed east–west distribution of the main genetic groups but with contrasting spatial genetic diversity. The spatio-temporal diffusion analyses showed that between 1–0.8 Mya and 570 Kya the lineage leading to C. fothergillii diverged from C. uniflora and arrived to the islands. Climatic niche projections hindcasted range expansions during glaciations, and contractions during the interglacial periods. Comparisons of climatic niches between the two study species indicated that temperature variables show evidence of niche conservatism while precipitation regimes supported niche divergence, even when considering the background environmental divergence. Our study indicates that glacial fluctuations affected the mainland/islands connections favouring speciation mediated not only by isolation, but also by climatic niche differentiation.
... Tests of the direct relationship between pollen color and both maximum temperature and UV show that temperature is the most likely abiotic factor to contribute to pollen-color variation. Similar longitudinal but not latitudinal variation has been observed for petal pigmentation in Calceolaria uniflora (Masc o et al., 2004) and body color in Drosophila americana (Wittkopp et al., 2011). While the drivers are not completely understood in those systems, correlative results in C. americana suggest that pollen color may be locally adapted to temperature, which has been shown to elevate flavonoid production (Winkel-Shirley, 2002). ...
Article
Full-text available
The evolution of flower color, especially petal pigmentation, has received substantial attention. Less understood is the evolutionary ecology of pollen pigmentation, though it varies among and within species and its biochemical properties affect pollen viability. We characterize the distribution of pollen color across 24 populations of the North American herb Campanula americana, and assess the degree to which this variation is genetically based. We identify abiotic factors that covary with pollen color and test whether germination of light and dark pollen is differentially affected by variable temperature and UV. Pollen color varies from white to deep purple in C. americana and is genetically determined. There was a longitudinal cline whereby pollen was darkest in western populations. Accounting for latitudinal variation, western populations experience elevated temperature and UV irradiance. Germination of light-colored pollen was reduced by 60% under high temperature, but dark pollen was unaffected. Exposure to UV reduced germination of light and dark pollen similarly. The cline in pollen color across the range may reflect adaptation to heat stress. This study supports thermal tolerance as a novel function of pollen pigmentation and contributes to growing evidence that abiotic factors can drive floral diversity.
... Morphological discontinuities will usually manifest as distinct clusters in phenetic analyses and can then be recognised at various taxonomic levels (Cron et al. 2007;Ereshefsky 2001;Lewis 1972;Mascό et al. 2004;Sebola and Balkwill 2009;Sneath and Sokal 1973;Sokal and Sneath 1963;Stebbins 1967). It has been recommended that when there are three or more diagnostic characters, taxa are recognised as species; discontinuity in two characters coupled with geographical disjunction is recognised at subspecific level; and a single diagnostic character with or without geographical isolation can be recognised at varietal level (Brysting and Elven 2000). ...
Article
Full-text available
The variation exhibited within three species of Barleria (B. bechuanensis, B. irritans and B. jubata) was studied to establish whether it was discrete or continuous. Morphological characters were examined and recorded in matrices. Cluster analysis was employed to impose a hierarchical non-overlapping association among operational taxonomic units (OTUs) while ordination was used to establish whether the variation was discrete or continuous. Discrete characters were determined from quantitative morphological data using box and whisker plots. Locality information for the OTUs was obtained from herbarium labels and used to generate maps to illustrate geographic distribution of taxa. Cluster analysis and ordination demonstrated that there was discrete variation within Barleria bechuanensis, B. irritans and B. jubata, which each split into two distinct clusters, although box and whisker plots illustrated that many quantitative characters overlapped within and between species. Since clear morphological gaps between clusters are assumed to be indicators of breaks in gene flow, the distinct clusters were recognised at species level.
... Variation in color and pattern are of the more vivid examples of morphological variability in nature. Taxa as diverse as spiders (De Busschere et al. 2012;Cotoras et al. 2016), insects (Katakura et al. 1994;Williams 2007), fish (Endler 1983;Houde 1987), amphibians and reptiles (Calsbeek et al. 2008;Allen et al. 2013;Balogová & Uhrin 2015;Rabbani et al. 2015), mammals (Hoekstra et al. 2006;Nekaris & Jaffe 2007) and plants (Clegg & Durbin 2000;Mascó et al. 2004) display natural variation in pigment or structural colorations. The distribution of colors in specific patterns play an important role in mate preference (Endler 1983;Kronforst et al. 2006), thermal regulation (Forsman et al. 2002), ...
Article
The use of image data to quantify, study and compare variation in the colors and patterns of organisms requires the alignment of images to establish homology, followed by color-based segmentation of images. Here we describe an R package for image alignment and segmentation that has applications to quantify color patterns in a wide range of organisms. 2.patternize is an R package that quantifies variation in color patterns obtained from image data. patternize first defines homology between pattern positions across specimens either through manually placed homologous landmarks or automated image registration. Pattern identification is performed by categorizing the distribution of colors using an RGB threshold, k-means clustering or watershed transformation. 3.We demonstrate that patternize can be used for quantification of the color patterns in a variety of organisms by analyzing image data for butterflies, guppies, spiders and salamanders. Image data can be compared between sets of specimens, visualized as heatmaps and analyzed using principal component analysis (PCA). 4.patternize has potential applications for fine scale quantification of color pattern phenotypes in population comparisons, genetic association studies and investigating the basis of color pattern variation across a wide range of organisms. This article is protected by copyright. All rights reserved.
Article
This study highlights the importance of flower color variation and attraction as a mechanism for pollination and protection of floral parts. As part of this study, a survey relating to flower color variation and differences in spotting pattern (nectar guides) was conducted on Rhododendron arboreum , a widespread tree species in the mountainous region of Uttarakhand state, at 43 different altitudinal locations. Seven original color morphs of flowers and five types of spot variation in the nectar guide were observed. The study underlines the role of flower color polymorphism in both pollination and adaptation to varied environmental conditions. Further, the significance of nectar guides in directing the visitor to the reward is discussed. This study has the potential to enhance existing knowledge about flower color variation and attraction to the environment.
Preprint
Full-text available
The use of image data to quantify, study and compare variation in the colors and patterns of organisms requires the alignment of images to establish homology, followed by color-based segmentation of images. Here we describe an R package for image alignment and segmentation that has applications to quantify color patterns in a wide range of organisms. patternize is an R package that quantifies variation in color patterns obtained from image data. patternize first defines homology between pattern positions across specimens either through manually placed homologous landmarks or automated image registration. Pattern identification is performed by categorizing the distribution of colors using an RGB threshold, k-means clustering or watershed transformation. We demonstrate that patternize can be used for quantification of the color patterns in a variety of organisms by analyzing image data for butterflies, guppies, spiders and salamanders. Image data can be compared between sets of specimens, visualized as heatmaps and analyzed using principal component analysis (PCA). patternize has potential applications for fine scale quantification of color pattern phenotypes in population comparisons, genetic association studies and investigating the basis of color pattern variation across a wide range of organisms.
Article
Background and aims: Studies of phenotypic plasticity in plants have mainly focused on (1) the effect of environmental variation on whole-plant traits related to the number of modules rather than on (2) the phenotypic consequences of environmental variation in traits of individual modules. Since environmental and developmental factors can produce changes in traits related to the mating system, this study used the second approach to investigate whether within-individual variation in herkogamy-related traits is affected by the environment during plant development in two populations of Datura stramonium , an annual herb with a hypothesized persistent mixed mating system, and to determine which morphological traits may promote self-fertilization. Methods: Full-sib families of two Mexican populations of D. stramonium , with contrasting ecological histories, were grown under low, mid and high nutrient availability to investigate the effects of genetic, environmental and within-plant flower position on flower size, corolla, stamen and pistil lengths, and herkogamy. Key results: Populations showed differences in familial variation, plasticity and familial differences in plasticity in most floral traits analysed. In one population (Ticumán), the effect of flower position on trait variation varied among families, whereas in the other (Pedregal) the effect of flower position interacted with the nutrient environment. Flower size varied with the position of flowers, but in the opposite direction between populations in low nutrients; a systematic within-plant trend of reduction in flower size, pistil length and herkogamy with flower position increased the probability of self-fertilization in the Pedregal population. Conclusions: Besides genetic variation in floral traits between and within populations, environmental variation affects phenotypic floral trait values at the whole-plant level, as well as among flower positions. The interaction between flower position and nutrient environment can affect the plant's mating system, and this differs between populations. Thus, reductions in herkogamy with flower positions may be expected in environments with either low pollinator abundance or low nutrients.
Article
Full-text available
Delimiting species is an important, but frequently difficult aspect of systematics that should be addressed using data from multiple sources. Here we combine morphometric analyses and environmental characteristics to delimit species in the South Andean and Patagonian taxonomically difficult species-group composed by Mulinum spinosum, M. echinus and M. leptacanthum (Apiaceae-Azorelloideae). Molecular phylogenies have shown that these three species form part of a polytomy together with other Mulinum species, and therefore these data are not useful for their delimitation. We include measurements of 25 morphometric variables from 163 herbarium specimens and perform univariate and multivariate principal component analysis (PCA) and discriminant analysis (DA) to establish the limits amongst the three mostly sympatric, morphologically similar, and phylogenetically unresolved species. We also use 19 bioclimatic and three geographic variables from localities of the specimens to infer environmental characteristics of the taxa and test their relation with morphological variation. Morphological evidence supports the inclusion of M. echinus within the morphologically and climatically variable M. spinosum, and rejects its recognition as a distinct taxon at any rank. On the contrary, M. leptacanthum is considered a morphologically distinct species, generally restricted to high altitude areas of the southern Andes with a cooler and wetter climate. Within the widespread M. spinosum, environmental gradients of precipitation and temperature relate to morphological gradients (e.g., in leaf and inflorescence sizes, in leaf acicularity, and in fruit-wing width). These last results showed that the large morphological variation in vegetative and reproductive characteristics of this species that grows in arid and semiarid habitats are related to regional climatic gradients that have probably been important in the evolution of this species' plasticity, diversification, and differentiation.
Article
Full-text available
This study compares the effect of ecological factors on phenotype plasticity in the natural populations of Acinos graveolens in Iran. In total, twenty six quantitative and qualitative morphological traits were examined in the sixteen populations at two sections: inter- and intrapopulation. In the interpopulation section, averages amounts of morphological characteristics of populations were examined, while in the intrapopulation, morphological features of the seventy six plant samples of the populations were investigated. The obtained data showed that these characteristics differed between and within the populations and the analysis of variance test and one-sample test have shown significant differences of all quantitative characteristics. The biplot Principal Correspondence Analysis (PCA-biplot) of traits showed some populations had indicator morphological characteristic(s) which were useful in detecting of these populations. Significant correlations were found between some quantitative characteristics with ecological factors. The studied populations were separated from each other in the PCA and Principal Coordinate Analysis (PCO) plots and the Unweighted Paired Group using Average method (UPGMA) tree. The PCA plot of morphological features showed that some characteristics separated from others and the rest ones were together and had overlapping. The individuals of the studied populations were separated from others in the PCA and PCO plots and the UPGMA tree of the morphological characteristics. Individuals of some populations arranged separately and placed far from others, but this condition was reverse in the rest populations and most of individuals placed near each other. The obtained data confirmed high variations within some populations.
Article
Full-text available
Nectar is a reward regularly offered by flowers to their pollinating birds. Studies on pollination by birds have focused on nectar as a reward while alternative reward systems have been poorly studied. Previous reviews have shown the presence of food tissues that serve as nutritive rewards to different classes of pollinators. However, in a recent revision of plant-bird interactions, the existence of a solid reward for birds has been disregarded. In South America at least three cases of this very particular bird-flower interaction have been recently studied. In this paper we describe the syndrome features (flower morphology, regard properties, position of the pollination unit, flower colour display and common visitors), that deviate from the typical ornithophilous syndrome, and its possible origin is discussed. Key words: flower, food reward, pollination, birds, syndrome, review.
Article
Full-text available
Una descripción de las principales Unidades de Paisaje y Areas Ecologicas de Santa Cruz y Tierra del Fuego
Article
Full-text available
Based on the measurements of 1063 flower reflection spectra, we show that flower colours fall into distinct clusters in the colour space of a bee. It is demonstrated that this clustering is caused by a limited variability in the floral spectral reflectance curves. There are as few as 10 distinct types of such curves, five of which constitute 85% of all measurements. UV reflections are less frequent and always lower in intensity than reflections in other parts of the spectrum. A further cluster of colour loci is formed in the centre of the colour space. It contains the colour loci of green leaves, several other background materials and only very few flowers. We propose a system to classify the reflection functions of flowers, and a set of colour names for bee colours.
Article
Full-text available
A new method of reconstructing the images of flowers as insects may see them is described. The method is based on the analysis of video recordings of flowers made through specially selected filters and a UV-sensitive camera. The recordings can be made under natural illumination conditions in the field. To select an appropriate set of filters we analyzed more than 1000 reflectance spectra of flowers. To reconstruct the images we used the results of optical studies of the honeybee eye and receptor spectral sensitivities of this insect. As an example of the application of the method, we show how the eye of the honeybee resolves color patterns and the corolla shape of a single flower of Orchis caspia Trautv.
Article
This study investigated fertility selection on a flower petal pigmentation polymorphism in Clarkia gracilis ssp. sonomensis. Natural populations are typically composed of nearly 100% spotted-petal plants, although rare populations contain a majority of unspotted plants. I compared fitness values for the two morphs using a simple fertility model to estimate selection for experimental arrays of plants placed into existing populations of different phenotypic frequencies. Both male and female reproductive success were estimated as well as the pattern of mating among phenotypes. Although the separate fitness components varied from no differences to a strong advantage for spotted plants, for every situation the selection calculations predicted an increase in the frequency of the spotted allele. Pollinator behavior and postpollination mechanisms may be responsible for the fitness differences. The apparent inability of the unspotted allele to spread though most natural populations is consistent with its selective disadvantage in this study.