Printer Friendly

The influence of dark central florets on insect attraction and fruit production in Queen Anne's Lace (Daucus carota L.)


The floral displays of flowering plants have numerous adaptations for attracting pollinators (e.g., Meeuse, 1961; Stephenson, 1979; Beach, 1982; Snow, 1982; McDade and Davidar, 1984; Weiss, 1991). Here, we focus on the importance of the colors displayed in umbels of Queen Anne's Lace (Daucus carota L.: Umbelliferae). Umbels of this species contain tens to thousands of small, bisexual florets, most of which are white. However, most umbels also have a few centrally located dark florets that create a color contrast. Such color contrasts may serve as long-distance signals that attract insects (Mulligan and Kevan, 1973).

In one of his works on pollination biology, Darwin (1888:8) stated that the central florets of Queen Anne's Lace ". . . are of no functional importance . . .", but did not present data to support this conclusion. In contrast, Eisikowitch (1980) found that bicolored umbels attract newly eclosed house flies (Musca domestica) much more strongly than unicolored umbels. In addition, he found that such flies were more often attracted to white circles of filter paper when dead flies had been attached, and that the strength of attraction was positively related to the number of flies attached. On the basis of these findings, Eisikowitch suggested that the central florets of Queen Anne's Lace evolved to mimic a visiting insect. His work, however, was done in a laboratory setting with a single insect species.

In this study, we sought to determine whether the central florets of Queen Anne's Lace attract insects in natural settings, i.e., with diverse insect taxa, and whether this influences fruit production. A similar study by Daumann (1973) compared insect visits to bicolored and natural unicolored (all white) umbels and concluded that the red central florets were not insect attractants. Here, we compare visitation by seven insect taxa between plants bearing the standard bicolored umbels and plants in which the central florets had been experimentally removed.


Reproductive biology of Daucus carota. - Queen Anne's Lace (wild carrot) is a naturalized, monocarpic weed (usually biennial) that grows abundantly in open fields and along roadsides in North America. Introduced from Europe, the species has successfully colonized the eastern regions of the U.S. and Canada. Anthesis occurs from June through September. Flowers on the terminal umbel open synchronously; the terminal umbel is the largest, accounting for ca. 50% of total fruit production. Secondary umbels open immediately after the terminal umbel has been pollinated. Following this, smaller tertiary and quaternary umbels open. The flowering period overlaps within and between plants, so a population is visited by insects for about 2 mo.

Most of the florets of Queen Anne's Lace exhibit the typical attributes of "small-fly syndrome" (Faegri and van der Pijl, 1966), including small size, radial symmetry, and easily reached nectaries, and thus are accessible to many insects. The most common visitors are flies and beetles, probably because the umbels have a slightly pungent scent. The florets are largely protandrous, so most are cross-pollinated (Fryxell, 1957; Owens, 1974).

Both the color and the size of the central cluster of florets can vary (Dahlgren, 1957). Size differences arise due to the number of florets that produce anthocyanins; typically, only 1-5 florets become dark. Color differences can be attributed to the time at which florets begin to produce anthocyanins (McClintock, 1967). When pigment production begins early in development, the color of a floret is darker. Some umbels lack a central spot altogether because anthocyanins are produced very late. A single plant sometimes contains both unicolored and bicolored umbels, but most ([greater than]80%) of the umbels are bicolored. In most populations the predominant phenotype is bicolored, but some populations of Daucus carota are composed of plants with entirely white umbels (Westmoreland, pers. observ.).

Study sites. - We monitored insect visitation and fruit production of 124 plants during the summers of 1988, 1989 and 1992. In 1988, we studied 16 plants on a hillside at The West Virginia Institute of Technology in Montgomery, West Virginia; in 1989, 16 plants in a plot adjacent to Emory University in Atlanta, Georgia, were used. Three study sites were used in 1992-21 plants were observed along roadsides in Hardin County, Iowa; 34 plants were studied in an open field in Jefferson County, Kentucky, and 37 plants were observed in Rockland County, New York.

Experimental design. - Before anthesis, plants were randomly assigned to either a control group or to a treatment group in which the red central florets were removed by clipping the pedicels that supported them. This experimental design was less than ideal because the treatment and control umbels differed in two ways: (1) color and (2) the presence of florets in the center of the umbel. Removal of the florets left a small (3-8 mm) open space in the umbel that was not conspicuous to the human eye. We considered removing an equal number of florets from control umbels, but reasoned that this would not be an improvement because treatment umbels still would have differed from control umbels in two ways: (1) color and (2) the location in an umbel where florets had been removed. Given that an ideal design could not be attained while maintaining the sample size calculated to achieve significant differences (i.e., we had to count [greater than]250,000 fruits to discriminate between two treatments), we chose to use a control that mimicked natural plants as closely as possible.

Field procedure. - All umbels on a plant received the same treatment. Control plants occasionally produced an umbel in which central florets were absent; such umbels were artificially given a central spot by applying red fingernail polish to a few centrally located white florets. Fewer than 5% of control umbels had to be modified in this way.

Umbels were individually marked by hanging 5 x 1-cm white cotton tags on the main stem where the pedicel emerged. These tags were 10-50 cm away from the floral display. We did not record data on tertiary and quaternary umbels, which contribute little to seed production because they are small in diameter and few in number.

Before the experiment began, umbels of neighboring plants were visited three times a day (1000, 1500 and 1900 h) to determine the period of maximum insect abundance. At the sites in Iowa, New York, West Virginia and Georgia, this was at 1500; at the Kentucky site abundance peaked around 1000. After anthesis began, plants involved in the experiment were visited once every 2nd day to record the number and taxa (usually to family level) of insects present. Eighty percent of observations were made during peak insect abundance and 10% in each of the other two time blocks. During each visit, the insects on top of each umbel (i.e., in a position to pollinate florets) were counted. Observation time per umbel varied from 30-60 sec, depending on how many insects had to be counted. Insects that arrived during observation were included in the count. Such arrivals were uncommon, occurring in [less than]5% of observations. At the beginning of the study, we collected the most common insect visitors (those which comprised [greater than]5% of visitors) for identification; composition of the visitor communities changed over time, however, so some of the early taxa were rare overall. Our observations continued until petals dropped from the flowers (2 wk for most plants), indicating the end of anthesis.

When petals dehisce and fruits begin to form, the umbels assume a flattened shape that allows accurate determination of diameter. We measured each umbel at its widest point, and used the diameters to calculate umbel area. Umbels were collected and frozen ca. 2 wk after fruits began to form. At this time, the rays had curled upward so that the umbels assumed the typical "bird's-nest" shape (Lacey, 1982; Heywood, 1983), but fruits had not dehisced. Fruit production was determined by counting the number of fruits [greater than]3 mm long. Most of the fruits [less than]3 mm had not enlarged after anthesis and thus probably did not contain viable seeds. The fruits of Daucus carota are schizocarps, which dehisce into two single-seeded mericarps (Lacey, 1982).

Data analysis. - Comparisons of fruit production/plant were done with analyses of covariance, using total umbel area/plant as the covariate. Because fruit production/plant was not normally distributed, we log-transformed the variable prior to analyses and back-transformed the means and standard errors for presentation in tables. Type I sums of squares were used and statistical significance for these analyses was set at 0.05.

Insect abundance was calculated as the number observed/[dm.sup.2] umbel area, averaged over all of the observation periods for a plant. This variable had a preponderance of low values, so we used Mann-Whitney U-tests for comparisons. Because these tests were based on smaller sample sizes, we set statistical significance at 0.1.


Fruit production. - Umbel area was strongly correlated with fruit production (N = 124, r = 0.69, P [less than] 0.01) and was the factor that explained most of the variation in the dependent variable (Table 1). Location effects were also significant. Plants in Georgia and West Virginia produced 25-150% more fruits than plants in the other locations (Table 2).

Removal of the central florets had a significant effect on fruit production, but the treatment by location interaction was also highly significant (Table 1). Post-hoc comparisons revealed that treatment had no effect on fruit production in Iowa, Kentucky, New York or West Virginia (Table 3). In Georgia, however, removal of the central florets was associated with a doubling of fruit production.
TABLE 1. - Analysis of covariance for fruit production

                       Fruit production per plant
Source                  df     MS            P

Umbel area               1    4.167    [less than]0.01
Location                 4    0.121    [less than]0.01
Treatment                1    0.096    [less than]0.05
Location*treatment       4    0.095    [less than]0.01
Error                  113    0.024

Variation among insect taxa. - Of the eight insect taxa, two showed a preference for control umbels when all locations were combined - mordelid beetles and syrphid flies [ILLUSTRATION FOR FIGURE 1 OMITTED]. Mordelids consistently preferred control umbels in the two study sites where they occurred. In Georgia, 67% of control plants were visited by mordelids, compared to 28.5% of treatment plants; in West Virginia, 40% of control plants attracted mordelids while no visits were recorded to treatment plants.

Preferences of other insect taxa varied from site-to-site (Table 4). Syrphid flies were present in three study sites, and showed a significant (P = 0.1) preference for control umbels in Iowa and New York, but no preference in Kentucky (P = 0.41). Chloropid flies occurred on four of the study sites, but showed a preference (P = 0.03) only in Kentucky.

Three taxa showed no treatment preferences in any location [ILLUSTRATION FOR FIGURE 1 OMITTED]: ants (Formicidae), which visited umbels in all five study sites; sciarid flies, which were present in Iowa, Kentucky and New York; and calliphorids in Iowa, Kentucky and New York. Miscellaneous taxa ("others") showed no preference in any location.

Variation among locations. - Despite considerable differences in composition of the floral and insect communities, we found that in each study site, a single insect taxon preferred control umbels over treatment umbels. In four of the five study sites, this taxon was rare, comprising [less than]5% of insect visitors (Table 5). The only exception was Kentucky, where chloropid flies preferred control umbels and comprised 28% of insect visitors.


Darwin (1888) believed that the central florets of Queen Anne's Lace have no function in attracting insects, and Daumann's (1973) finding that unicolored umbels receive as many insect visits as bicolored umbels seems to support the hypothesis. In this study, however, we found that the central florets do influence insect attraction. Relatively few taxa responded to our experimental manipulations of the umbel display, but those that did always favored plants with bicolored umbels. The findings of Eisikowitch (1980) also suggest that some taxa prefer bicolored umbels.
TABLE 2. - Least-squares means(a) [+ or -] SE of fruit production
per plant in the different study sites. Means with the same letter
are not significantly (P [less than] 0.05) different

Location           n         Fruit production

Georgia            16      3639 [+ or -] 444(A)
Iowa               21      1884 [+ or -] 180(BC)
Kentucky           34      1741 [+ or -] 141(B)
New York           37      2338 [+ or -] 145(BC)
West Virginia      16      3147 [+ or -] 427(AC)

a Based on ANCOVA with effects of umbel area removed


A response to the presence of red florets is puzzling because the visual range of most insects extends from 300-650 nm, which excludes red (Burkhardt, 1964). Insects may see the central florets as a UV-reflecting spot against a background of UV-absorbing white florets; alternatively, the central florets may appear to them as a dark spot against a white background (Eisikowitch, 1980). Published photographs (Mulligan and Kevan, 1973) show that the white florets of Queen Anne's Lace absorb UV and reflect the other insect-visible wavelengths, but the spectral properties of the central florets are unknown (P. G. Kevan, pets. comm.).

Even when a taxon preferred control umbels, higher visitation did not lead to greater fruit production. In Iowa and New York, syrphid flies showed a strong preference for control umbels, but made up a small component (1-4%) of the visitor community. Control umbels in Kentucky were favored by chloropids, which comprised 28% of the insect community. [TABULAR DATA FOR TABLE 4 OMITTED] This preference even by a large component of the visitor community did not translate into higher fruit set (Table 2). This suggests that for the central florets to be adaptive, the attracted insect must be very common, or must be a very effective pollinator. It is important to note, however, that our quantification of reproductive success focused exclusively on fruit set, which is the female component of reproduction. If the central florets benefitted the male component, pollen donation, we would not have detected it. For some plants, the [TABULAR DATA FOR TABLE 5 OMITTED] color of the floral display has a stronger influence on pollen donation than on fruit production (Stanton et al., 1986).

Removal of the central florets in Georgia actually led to an increase in fruit production (Table 3); this could be explained if mordelid beeries, which preferred control umbels at that site, fed upon ovules instead of nectar or pollen. To examine this, we correlated fruit production/[dm.sup.2] with mordelid visitation/[dm.sup.2] for Georgia plants that were visited by the beetles. The correlation is negative and nonsignificant (N = 8, Spearman-rank r = -0.381, P [greater than] 0.1). Control umbels in West Virginia were also favored by mordelid beetles, but did not suffer decreased fruit production. Thus, it seems that the paradoxical result observed in Georgia cannot be explained by differences in insect visitation.

Two taxa (Syrphidae and Chloropidae) varied in their preferences from site to site. This phenomenon has been reported in other studies, and may explain why Daumann's (1973) results contrasted with ours. In studying pollination of Raphanus raphanistrum, Kay (1976) found that syrphids (Eristalis spp.) consistently favored one floral morph in four study sites and showed no preference in a fifth. In another study of the same species Kay (1978) reported that the bumblebee Bombus terrestris switched preferences between study sites, and honeybees Apis mellifera showed preferences in some locations and no preference in others. Furthermore, the preference of an individual insect often changes from day to day (Kay, 1982). If preference-switching is a common feature of pollination biology, studies done in a single location may be insufficient to evaluate the adaptive significance of a floral trait.

Taken together, our findings show that the effect of the central florets on fruit production varies spatially and, perhaps, temporally. Because the vast majority of insects on our study sites did not show a preference for bicolored umbels, the preference shown by a few species had no substantial influence on fruit production. Queen Anne's Lace is a naturalized species in North America, and thus the insects that we observed may not be representative of the pollinators in its native European habitat. The scarcity of insects that favor bicolored umbels could explain the existence of populations entirely composed of unicolored umbels.

Acknowledgments. - We thank Gail Westmoreland and John Parnell for assistance with field work. This manuscript was improved by discussions with the population biology, ecology and evolution study group at Emory University, and by two anonymous reviewers.


BEACH, J. 1982. Beetle pollination of Cyclanthus bipartitus (Cyclanthaceae). Am. J. Bot., 69:1074-1081.

BURKHARDT, D. 1964. Colour discrimination in insects. Adv. Ins. Phys., 2:131-173.

DAHLGREN, K. V. O. 1957. De morkroda mittblommorna hos vild morot. Sven. Bot. Tidskr., 51:581-604.

DARWIN, C. 1888. The different forms of flowers on plants of the same species. University of Chicago Press, London. 352 p.

DAUMANN, E. 1973. Uber die vermeintliche Bedeutung der Zentralblute(n) in der Dolde von Daucus carota L. fur die Bestaubungsokologie und als Schultz vor Weidertieren. Preslia, Praha, 45: 320-326.

EISIKOWITCH, D. 1980. The role of dark flowers in the pollination of certain Umbelliferae. J. Nat. Hist., 14:737-742.

FAEGRI, K. AND L. VAN DER PIJL. 1966. Principles of pollination biology. Pergamon Press, Oxford. 224 p.

FRYXELL, P. A. 1957. The mode of reproduction in higher plants. Bot. Rev., 23:135-233.

HEYWOOD, V. H. 1983. Relationships and evolution in the Daucus carota complex. Isra. J. Bot., 32:51-65.

KAY, O. K. N. 1976. Preferential pollination of yellow-flowered morphs of Raphanus raphanistrum by Pieris and Eristalis spp. Nature, 261:230-232.

-----. 1978. The role of preferential and assortative pollination in the maintenance of flower colour polymorphisms, p. 175-190. In: A. J. Richards (ed.). The pollination of flowers by insects. Academic Press, New York.

-----. 1982. Intraspecific discrimination by pollinators and its role in evolution, p. 9-28. In: A. J. Armstrong, J. M. Power and A. J. Richards (eds.). Pollination and evolution. Royal Botanic Gardens, Sydney.

LACEY, E. P. 1982. Timing of seed dispersal in Daucus carota. Oikos, 39:83-91.

MCCLINTOCK, B. 1967. Genetic systems regulating gene expression during development, II. The role of the nucleus. Dev. Biol. Suppl., 1:84-112.

MCDADE, L. A. AND P. DAVIDAR. 1984. Determinants of fruit and seed set in Pavonia dasypetala (Malvaceae). Oecologia, 64:61-67.

MEEUSE, B. J. D. 1961. The story of pollination. Ronald Press, New York. 243 p.

MULLIGAN, G. A. AND P. G. KEVAN. 1973. Color, brightness, and other floral characteristics attracting insects to the blossoms of some Canadian weeds. Can. J. Bot., 51:1939-1952.

OWENS, S. J. 1974. An examination of the floral biology, breeding system and cytology in species of the genus Daucus and related general in the tribe Caucalidaeae (Umbelliferae). Ph.D. Dissertation Univ. of Reading. 119 p.

SNOW, A. A. 1982. Pollination intensity and potential seed set in Passiflora nitifolia. Oecologia, 55:231-237.

STANTON, M. L., A. A. SNOW AND S. N. HANDEL. 1986. Floral evolution: attractiveness to pollinators increases male fitness. Science, 232:1625-1627.

STEPHENSON, A. G. 1979. An evolutionary examination of the floral display of Catalpa speciosa. Evolution, 33:1200-1209.

WEISS, M. R. 1991. Floral colour changes as cues for pollinators. Nature, 354:227-229.
COPYRIGHT 1996 University of Notre Dame, Department of Biological Sciences
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1996 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Westmoreland, David; Muntan, Chad
Publication:The American Midland Naturalist
Date:Jan 1, 1996
Previous Article:Recent fire history of the northern Great Plains.
Next Article:An anther-smut disease (Ustilago violacea) of fire-pink (Silene virginica): its biology and relationship to the anther-smut disease of white campion...

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters