Great Basin Naturalist
Volume 52 | Number 2
Article 5
9-22-1992
Pollinator preferences for yellow, orange, and red
flowers of Mimulus verbenaceus and M. cardinalis
Paul K. Vickery Jr.
University of Utah
Follow this and additional works at: http://scholarsarchive.byu.edu/gbn
Recommended Citation
Vickery, Paul K. Jr. (1992) "Pollinator preferences for yellow, orange, and red flowers of Mimulus verbenaceus and M. cardinalis," Great
Basin Naturalist: Vol. 52: No. 2, Article 5.
Available at: http://scholarsarchive.byu.edu/gbn/vol52/iss2/5
This Article is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been
accepted for inclusion in Great Basin Naturalist by an authorized administrator of BYU ScholarsArchive. For more information, please contact
[email protected].
C,.al Basin Naturalil' 52(2). pp.145-148
POLLINATOR PREFERE CES FOR YELLOW, ORANGE, AND RED FLOWERS OF
MlMULUS VERBENACEUS AND M. CARDINALlS
Robert K. Vickery, Jr,l
ASSTRAcr.-Red, orange, and yellow morphs of Mimulm verbenaceus and M. cardinalts were field tested for pollinator
preferences. The species are closely similar except that M. verbenaceus flowers have partially re£lexed corolla lobes, whereas
M.. cardinali:> flowers have fully reflexed corolla lobes. On the basis of over 6000 bumblebee and hummingbird visits, highly
significant (p < .(01) patterns emerged. Yellow, which is the mutant color morph in both species and is determined by a
single pair of genes. was strongly preferred by bumblebees and strongly eskewed by hummingbirds in both species. Orange
and. to a lesser extent, red were strongly preferred by hummingbirds but eskewed by bumblebees in both species. Thus,
strong, but partial, reproductive isolation was obselVed betvleen the yellow mutants and the orange- to red-flowered
popurations from which they were derived. Color-yellOW" versus orange and red-appeared more important than
shape-partially reflexed versus fully reflexed corolla lobes-in determining the preferences of the guild of pollinators in
this particular test environment for Mimulus verbena-cem and M. cardinalis.
KEy wonls: Mimulus, speciation,flower colors, poUituttorpreferences, bumblebee.s, hummingbirds.
How much ofachange in flower color and/or
shape is enough to lead to a change in pollinators
and hence to reproductive isolation and potentially to speciation? The flower color and shape
morphs ofMimulusverbefUlceus Greene andM.
cardinalis Douglas prOvide an excellent system
for addressing this intriguing question.
anthocyanin pigments in the corolla lobes.
Thus, parallel series of red, orange, and yellow
color forms are available for both M. verben= and M. carWnalis (Table 1).
Mimtdus verbenaceus and M. cardinaiis are
similar, closely related species in section
Erythranthe (Grant 1924); however, their flowel~
MATERIALS
Mirrwlus verbenaceus and M. carWnalUi are
typically hright red flowered and humminghird
pollinated. However, yellow-flowered morphs
occur in M. verbenaceus, e.g., in a population at
Vassey's Paradise, Grand Canyon, Arizona, and
in M. cardinalis populations, e.g., on Cedros
Island, Baja California, Mexico, and in the
Siskyou Mountains, Oregon. My experimental
hyhridizations show that yellow is due to a Single
pair of recessive genes that limit the floral
anthocyanins to small dots in the corolla throat.
Intermediate, orange-flowered forms are
known in M. vet'benaceus, specifically thejopulation at Yecora, Sonora, Mexico. An , an
intermediate, orange-flowered form of M. car-
dinalis was obtained by repeated cycles ofselection. In both cases orange is due to a single pair
of quantitative genes that reduce the amount of
differ in shape. Those of M. verbenaceus
have only the upper pai r ofcorolla lobes sharply
reflexed, giving the flowers a partially tubular
aspect. The side pair of lobes and the labellum
cUlVe gently fotward forming a bee landing platform. Flowers of M. canlinaiis have both the
upper and side corolla lobes sharply reflexed,
giving the flowers a fully tubular shape. The
labellum is thrust forward and is folded on itself
forming a ridge instead of a landing platform.
Shapes of the flowers of both species would
seem to invite hummingbirds. Flowers of M.
verbenaceus but not those of M. cardinalis
would appear adapted for bees as well. However, flowers of all three color morphs of both
species showed no reflectance patterns in the
ultraviolet, that is, no putative bee nectar
guides. Thus, flower shapes of M. verbenacetls
and M. cardinalis are sim~ar in some respects
but differ in others of potential significance to
pollinators.
1De~rlment of BiOlogy, Ul'liYe~ity of Utah, Salt Lake aty. Utah 84112.
145
146
GREAT BASIN NATURALIST
PLAN
TABLE
[Volume 52
1. Origin of populations studied.
Mimulus verbenaceus Greene
Vasseys Paradise, Grand Canyon, Arizona, USA, elev.
--650 m
Red IDOJph '" culture number 14,088
YeUCl\lImorph '" culture number 14,089
Yecora, Sonora, Mexico, elev. -L55O m
Orange:::: culture number 13.256
The effect of flower color and flower shape
on pollinator preferences will be addressed
stepwise. First, pollinator preferences for
color-red, orange, and yellow-·will be ascertained for M. veroe,uu;eus plants only, holding
flower shape constant. Second, red-, orange-,
Mimulus cardinam Douglas
and yellow-flowered M. cardinalis plants will be Isla- Cedros, Baja California, Mexico, elev. -100 m
added to the experiment. Are pollinator preferRed morph '" culture number 13,106
Yellow morph '" culture number 13,250
ences for red, orange, and yellow flowers of M.
Orange:::; culture number 13,249
cardinalis the same as for those of the M. ver(obtained by selection from the red morph)
benaceus series? Note that the pigments are
identical (Vickery 1978). Or, does the difference
in corolla shape between the two species lead to . benaceus plants for the second partoftheexpera difference in pollinator preferences?
iment.
Pollinator visits to the flowers were observed
and
recorded
for
an average of 1\02 hours per
METHODS
observation period for 15 periods for each ofthe
two parts of the experiment (Tables 2, 3). Time
Seeds for each of the six populations of the
of day of the observations was varied to be sure
study (Table 1) were collected in the wild or
haJVested from transplants brought into the of noting all the different kinds of visitors. To
greenhouse except those of orange M. car- count as a visit, a hummingbird had to thrust its
dinalis, which were obtained by selection. A beak into a flower. A bee had to land on the
large population of red M. cardinalis from flower and crawl into the flower far enough to
Cedros Island was grown and the most orange- brush the stigma and anthers. A fly, butterfly,
red flowered plant self-pollinated. Its progeny etc., had to walk on the reproductive structures.
included several orange-flowered plants. Prog- The numbers of flowers rather than plants of
eny of these plants were grown and found to each color of each species were recorded for
breed true for orange and were used as the each observation period.
.
source of seeds for the orange M. cardinalis
For analysis of visits, chi-square tests were
morpho
run for each observation period for each part of
Seeds of the six populations were sown in
the experiment. The null hypothesis was that
early April 1988 in the University ofUtah greenhouse, following which seedlings were trans- hummingbirds or bumblebees (very few flies,
planted into 4" plastic pots and grown to butterflies, etc., visited the flowers and were not
flowering. Pots were placed in plastic trays to listed) would visit the three colors of flowers of
facilitate bottom-watering, plants being ran- M. veroenaceus in the first part of the experiment and the three colors of M. vero",uu;eus
domly arranged as to flower color.
When plants began flowering, they were and M. cardinalis in the second part of the
moved outdoors to test pollinator preferences. experiment in proportion to the numbers of
Instead of using Red Butte Canyon Natural those flowers in the experimental population
Research Area as before (Vickery 1990), with its (Tables 2, 3). If the overall chi-square value for
relative paucity of pollinators, 1 scattered the a period of, for example, bee visits to M. verplants on a lawn adjacent to native gambel oak benace". or hummingbird visits to M. cardinalis
clumps at the mouth of Parley's Canyon of the
indicated a significant deviation from e'1"'cted
Wasatch Mountains in an area rich in pollinavalues,
then
the
frequency
ofvisits
to
each
color
tors. Some 50 to 100 plants of each color morph
of M. veroenaceus made up the artificial popu- was inspected. Those high or low enough that
lation of the first part of the experiment. Some their term in the chi-square equation was large
50 to 100 plants of red and of orange M. car- enough by itself to produce a significant deviadinalis plus 20 plants ofyellow M. cat-dinalis (all tion at the 5% level were considered to be
that were available) were added to the M. ver- significant (Tables 2, 3).
1992]
147
MIMULUS POLUNATOR PREFERENCES
TAJ112. 2. Pollinator preferences for red. orange, or yeUOY.' flowers of Mimulw verberuu:eus in 1988.
Bumblebee visits
Nllmbers of flowers
Month:day:time
H6,1630
7,29,0745
7,30,0710
8,02,1640
8,03,0630
8,03,1540
8,04,0640
8,%0715
M5j645
8,05,1830
8,06,()840
8,0&1445
M6,181O
8,07,1515
8,08,0725
Red
Orange
Yellow
48
56
48
56
91
79
77
101
117
73
100
104
104
88
98
117
119
91
70
74
114
74
133
172
178
169
174
174
151
150
142
142
124
85
92
120
86
120
126
126
126
126
130
130
118
Hummingbird visits
Red Orange
Yellow
28.1.:l
30
1981 <.001
58 <.200
67 <.010
53 <.001
81 <.001
209t <.001
52t <.001
125 <.001
126t <.001
73t <.001
29li <.001
oot <.001
257t <.001
68t <.001
13li <.001
741
&lSO
523t
SO
67
88t
99t
74
5
71
221
41
159t
&l105
01
4!
24
31
53
311
01
3,31
121
51
121
32
P
Red
°°
55t
27
33
1001
83
91
381
75
66
49
31
52
32
Orange Yellcm
3
8li
66
49
79t
24lt
145t
77t
149;
lSOt
lOOt
94t
48t
125t
67t
°
29
70
27
361
183
170
281
921
821
2&l-
24.
11
51
51
P
<.100
<.010
<.OlD
<.010
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
'1 or J. "signifx:=tly high ,X' kJW; see ten.
TABLE:3. Pollinator preferences for red, orange, or yellow flowers of M.
Number of flowers
Month:day:time
Red
Orange
Yellow
verheflllcettS
Bumblebee visits
Red Orange
and M. cardinalis in 1988.
Hummingbird visits
P
Red
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
23
70
17l
13
401
60
196
168
115
44
7l
54
7
21
66
40t
70t
167t
10
/lOt
sot
166t
147t
63t
3li
64t
37
2
JOt
72t
28
61
137
4
81
34
63
lJO
59
91
77
55
II
34
201
25
36
117t
YellQ\.\'
Orange Yellow
P
Mimulus verbenaceus
8,08,1600
8,09,0750
8,09,1705
8JO,0815
HlU640
8,IL081O
H2,0805
B,l2,17oo
8,13,0855
8,IH800
8,14,0815
8,15,0040
8,15,1700
8,16,0830
8,17,0630
117
115
115
145
145
175
200
200
180
180
212
184
184
206
214
92
7.3
73
90
90
83
111
111
83
87
81
94
94
112
86
132
116
116
143
143
177
198
198
175
175
165
183
183
153
177
17!l\
161
2.
841
131
841
961
51
541
31
271
391
21
141
31
29
36
211
129t
56
lOOt
97
91
66
62t
131t
124t
lOOt
202t
237t
160
120t
4!
lOOt
162t
167t
50t
174t
128t
37
36
3
81
21
172t <.IXH
<.001
<.001
<.001
<.001
<.001
<.001
31
181
1351
3
381
01
991
163
561
21
241
38
2
01
261
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.exH
<.001
<.001
<.010
<.300
<.001
<.001
Mimulus cardinalis
8,08,1600
8,09,0750
8,09,1705
8,10,0815
8,10,1640
8JL081O
8,12,0805
6,IH700
8,/3,0855
8,IH800
8,14,0815
8J5m4O
S,IH7oo
8,16,0830
8,17,0630
79
69
69
61
61
61
65
65
64
89
83
53
53
79
89
.,. or!" significantly higb nT low: see lert.
47
45
4,5
39
39
55
51
51
42
81
69
7l
7l
78
79
61
32
32
23
23
12
18
18
14
14
15
15
15
21
18
381
211
18
49
48
2n
33
18
35
28
39
20
13
181
191
59
37
8
59
101
62
26
8
40
15
24
33
4
21
47
89t
34t
22t
26
48t
49t
34t
10
17
16t
19t
26t
<.001
<.001
<.001
<.010
<.001
<.001
<.001
<.050
<.010
<.001
<.001
<.(XU
6t <.010
34
12
<.eX)!
<.001
6
53
20
95t
88
41
92
52
58
24
74t
1I5t
il <.001
01 <.001
241 <.001
°
°
<.100
°
<.010
121 <.010
<.020
13 <.001
21 <.100
01 <.010
14 <.300
3. <.020
01 <.001
01 <.001
41 <.001
148
[Volume 52
GREAT BASIN NATURAUST
RESULTS
Pollinators showed clear, very highly significant (p < .OOl) preference for or avoidance of
ye,llow flower color, but less clear preferences
for or avoidance of orange or red flower colors.
Bumblebees-principally B"mbus oppositus and
B. huntii-strongly preferred yellow in both M.
veroenaceus and M. cardinolis. Difference in
shape did not appear to matter. Hummingbirds-principally Selasphorus platycerusstrongly eskewed yellow in both species (Tables
2,3). Again, difference in shape did not appear
lo matter.
Hummingbirds significantly (p < .001) preferred orange M. veroenaceus flowers and
showed a tendency to prefer orange M. car,
dinalis flowers as well (Tables 2, 3). This preference for orange over red flowers should not have
been surprising in view of the fact that orange
and red are equally conspicuous to hummingbirds (Grant and Grant 1968, Raven 1972).
Strong preferences and aversions for yellow
are particularly interesting because yellow is the
mutant color in both species. So, a new yellow
mutant ofeither species would be preferentially
visited by bumblebees and preferentially
avoided by hummingbirds, but not in all-ornone reactions. Apparently then, with the species of pollinators tested, we are seeing the
estabUshment of real, but partial, reproductive
isolation due to the mutation of a single pair of
color genes.
LITERATURE CITED
GRANT, A. L. 1924. A monograph of the genus Mlrnulrl.s.
Annal.. of the Missouri Botanical Garden II: 99-389.
CUANT, K.. and v. CHANT. 1968. Hummingbi.rds and tbeir
flowers. Columbia University Press, Neo.v York. 115 pp.
RAVF:N, P. H. 1972. Why are bird-vi~ted flowers predominantly red? Evolution 26: 674.
VICKERY, R. K., JII. 1978. Case studies in the evolution of
species complexes in Mimulus. Evolutionary Biology
1I, 404-506.
_ _---;" 1990. Pollination experiments i.n the Mimillus
cardinalis-M. lewisii complex. Great Basin Naturalist
SO,155-159.
Received 21 October 1991
Accepted 1 May 1992