Wind Pollination In The Humboldt Bay Wallflower

Investigation of Wind Pollination in the Humboldt Bay Wallflower (Erysimum menziesii ssp. eurekense) at the Lanphere-Christensen Dunes Joshua Der, Sarah Jordan, Veronica Vega Senior Thesis in Biology Humboldt State University, Arcata, CA 95521 May 2003 Final report submitted to Humboldt Bay National Wildlife Refuge.

Abstract We examined wind pollination as a potential contributor to the pollination success of the Humboldt Bay wallflower. We selected 15 plants each on dune ridges and in dune hollows to test the effects of different wind regimes. Plants were emasculated and caged to exclude insect pollinators and ensure that wind was the only mechanism available for pollination. Stigmas were collected five days after emasculation and pollen loads were counted. We found no significant differences in the number of Erysimum or heterospecific pollen grains, or the ratio of each on the stigmas between ridges and hollows. There may not have been enough of a difference between treatments to affect pollen deposition. We found more pollen in both treatments than can be attributed to contamination from emasculation. This suggests that wind may play a role in the pollination system of Erysimum.

Wind Pollination in the Humboldt Bay Wallflower Introduction Plants use a variety of mechanisms to achieve pollination. These pollination systems are variable, cryptic, and often poorly understood in most plants. Plants often utilize more than one pollination system in order to ensure pollination success and increase fecundity (Whitehead, 1968; Goodwillie 1999). Ambophily is a pollination system that uses both insects and wind as vectors of pollen transfer (Culley et al., 2002). Ambophily has evolved in several unrelated families of plants including the Asteraceae, Brassicaceae, Ericaceae, Salixaceae and others (Culley et al., 2002 and references therein). Recent research suggests that ambophily is a transitional phase between exclusively animal and wind pollinated systems (Culley et al., 2002). Ambophily may be selected for by low insect visitation, high winds, and semelparity (Goodwillie, 1999). Wind pollination is generally considered to be a primitive pollination system, and the diversification of angiosperms has been associated with the evolution of insects as pollinators. However, morphological adaptations of flowers in some animal pollinated plants may promote wind pollination. Flattened or reflexed corolla lobes and exerted or elongating stigmas are adaptations which increase the likelihood of wind pollination (Goodwillie, 1999). The contribution of wind pollination to reproductive fitness varies widely between different plant groups. The proportion of the total pollen transferred by wind varies from 8% to as much as 93% (Culley et al., 2002). The Humboldt Bay wallflower (Erysimum menziesii ssp. eurekense, Brassicaceae) is a small Endangered herb endemic to the windy coastal foredunes near Humboldt Bay, California. This habitat faces increasing pressure from off road vehicle use and invasive plants. Erysimum produces a small inflorescence as early as mid-February, when pollinator abundance is low. The flowers are small and have many characteristics commonly associated with bee-pollinated plants: showy yellow flowers, abundant pollen, nectar rewards, and petals that provide a landing platform for bees. Erysimum also has flattened petals and an elongated style exposing the stigma, increasing the opportunity for wind pollination. Although insect visitation is limited, fruit and seed set are nearly complete (M. Mesler and J. Sawyer, unpublished). These factors suggest that Erysimum may utilize wind as a vector of pollination in addition to insect pollination.

2

Wind Pollination in the Humboldt Bay Wallflower Understanding the reproductive biology of this Endangered plant, allows us to develop effective management strategies to conserve the species and preserve the integrity of its rare community. Since there has been little research on ambophily to date, it is important to examine this as a potential mechanism of pollination in plants traditionally thought to be exclusively insect pollinated. The purpose of this study was to determine if Erysimum utilizes wind pollination as part of its pollination system. We expected to find a higher rate of pollination in exposed areas than in sheltered areas when we emasculated plants and excluded insect pollinators. Methods Study Species and Site Erysimum menziesii ssp. eurekense is a Federally and State Endangered species, endemic to the Humboldt Bay area. It is found only on the Lanphere and the Samoa dunes. It typically flowers between mid February and mid June, with peak flowering between March and April. Insect and self-pollination have both been observed in Erysimum. An average of 29 flowers are produced by an inflorescence over its lifetime and these flowers open acropetally at a rate of 2.2 flowers per day (M. Mesler and J. Sawyer, unpublished). Plants usually display inflorescences 10-25 cm tall, and larger plants sometimes branch. Plants are monocarpic and inflorescences bolt from a basal whorl of leaves after 1-2 years of vegetative growth. Erysimum’s most abundant insect pollinator is the solitary bee, Habropoda misirabilis (M. Mesler and J. Sawyer, unpublished). Other important pollinators include Bombus melanopygus, B. mixtus, B. vosnesenskii and Lasioglossum pavanotum. Occasionally ants and Elaterid beetles will visit the flowers. The Lanphere Dunes Preserve is a unit of the Humboldt Bay National Wildlife Refuge located west of Arcata in Northern California and is characterized by two general habitat types: foredunes and dune forest. The foredunes are characterized by a series of dune ridges and hollows. Erysimum occurs predominantly in exposed sandy areas on the foredunes in the northern section of the preserve and appears to have a clumped distribution (personal observation). The preserve is an important site that protects these

3

Wind Pollination in the Humboldt Bay Wallflower habitats and their rare vegetation from encroachment by invasive plants. Invasive plants have been actively and thoroughly eradicated from the preserve, providing a sharp contrast to the neighboring dune habitat. The surrounding areas invaded by yellow bush lupine and European beach grass no longer support Erysimum, but population demographic studies within the Lanphere Dunes Preserve have shown the population to be increasing (lambda=1.10) (E. Jules, personal communication). Pollen Trapping We collected airborne pollen using pollen traps. These were constructed from glass microscope slides, coated with a thin layer of Vaseline®, and mounted at flower level on wire supports (Kearns and Inouye, 1993). Four traps were placed in two locations: at 1 m and 5 m downwind from a flowering Erysimum in hollows and on ridge tops. We collected pollen twice on sunny days in late February, each for a 48-hour period. Slides were visually scanned at 40X with a compound scope to detect the presence of pollen. Sample Size and Plant Selection We selected 15 plants each from dune ridges and hollows. We chose these two locations because they seemed to experience relatively different wind regimes: windy and calm, respectively. Plants were selected based on four fundamental criteria: 1) plants with 7-15 unopened flowers remaining, 2) unbranched inflorescences, 3) were 0.25-10 m away from the nearest flowering Erysimum, and 4) were approximately the same size and age. Emasculation and Insect Exclosure We emasculated the unopened flowers from each plant in order to eliminate autonomous self-pollination. We also removed remaining pollen-bearing anthers to avoid geitonogamy. Immediately after emasculation we caged each plant. We constructed cages from 1/4 inch steel mesh hardware cloth and three stakes made from 1/4 inch welding rod. Cages were cylinders 0.3 m in diameter and 0.6 m tall. They were large enough that plants could sway freely without contacting the sides. Stakes extended

4

Wind Pollination in the Humboldt Bay Wallflower approximately 20-25 cm into the sand ensuring that cages would not be blown over when exposed to high winds. The mesh was fine enough to exclude most common insects, but would still allow wind to pass through the cage. Study Period and Pollen Analysis Plants were caged for five days beginning March 25-26, 2003, with ridge plants emasculated on March 25th and hollow plants on March 26th. March 25th was rainy; each subsequent day in the study period was clear with long periods of strong winds. At the conclusion of the five-day exposure period, we removed stigmas from all flowers that had been unopened at the start of the study period. We mounted stigmas on microscope slides in fuchsin gel, and set them with a coverslip using a lighter to heat and melt the gel. We visually compared pollen found on the collected stigmas with voucher Erysimum pollen collected from anthers and mounted in fuchsin gel. We also collected voucher pollen from three common plants flowering during this study (Fragaria, Salix, Solidago) to attempt identification of heterospecific pollen found on stigmas. We counted the number of Erysimum and heterospecific pollen grains on each stigma at 100X magnification. Emasculation Technique Calibration In order to evaluate the level of potential contamination due to our emasculation techniques we performed a series of control emasculations. Each member of the field team emasculated a series of three flowers. Each person emasculated a young flower bud, a ready to open flower, and a newly opened flower. After emasculation, we immediately removed the stigmas and mounted them on a microscope slide in fuchsin gel. We counted the number of pollen grains on these stigmas to evaluate the level of contamination each person was responsible for during emasculation. Data Analysis We calculated mean numbers of pollen grains for each plant to make plants the unit of observation. We compared the mean number of Erysimum and heterospecific pollen grains per flower for each plant between ridges and hollows using a t-test. We

5

Wind Pollination in the Humboldt Bay Wallflower also compared the mean proportions of Erysimum pollen found on stigmas from ridges and hollows using a t-test. The level of contamination was compared to the number of Erysimum pollen on stigmas in ridges and hollows with a t-test. Results Pollen traps collected very little pollen from any species. Pollen grains that were caught on the microscope slides were not Erysimum pollen. In our emasculation calibration procedure we found an average of 4.0 Erysimum pollen grains on each stigma. This was significantly lower than the number of Erysimum pollen grains found on treatment stigmas (p=0.01, df=35). The mean total number of pollen grains received on stigmas per plant ranged from 3.57 to 196.67 among plants in both ridges and hollows. Due to this wide range in pollen loads, the variance of our means for the total number of pollen grains was high – 2388 and 747 for ridges and hollows, respectively. The mean number of Erysimum, heterospecific, and total pollen grains as well as the proportion of pollen grains was not different between ridges and hollows (Table 1, Figure 1).

Table 1: Mean number of pollen grains per stigma ± standard error. P-value from a twotailed t-test. Ridge Hollow P-value

Erysimum

Heterospecific

Total

Ratio Erysimum:Heterospecific

13.9±4.3 11.0±4.2 0.64

38.8±8.6 29.5±5.1 0.36

52.7±12.6 40.6±7.1 0.41

0.33±0.04 0.56±0.2 0.31

6

Wind Pollination in the Humboldt Bay Wallflower

Mean Number of Pollen Grains

Number of Grains

70 60

Ridge

50

Hollow

40 30 20 10 0 Erysimum Pollen

Heterospecific Pollen

Total Pollen

Pollen Type Figure 1: Mean number of pollen grains found on stigmas on ridges and in hollows. Error bars indicate standard error. Differences were not significant.

Discussion The lack of pollen found on the pollen traps compared to the levels of pollen found on stigmas suggests that the pollen traps were not efficient at catching airborne pollen. Another possible explanation for the low pollen counts could be attributed to the method used to examine pollen grains. This method was different than that used to count pollen grains on stigmas. Because we didn’t stain the pollen grains on the traps, they were difficult to see, and since we were only able to look at the slides under 40X magnification, these pollen counts may be conservative. We chose to examine the pollen loads on the ridges and in the hollows because we expected to find a difference in the wind regimes and subsequently a difference in the contribution of wind to pollination among these two sites. As we spent time on the dunes, however, we observed that during strong winds, the ridges appeared to funnel the wind into the hollows. We did not directly measure wind speed and are unable to quantify any difference in wind regimes between our two treatments. Because we found no difference in the pollen loads between ridges and hollows, this may suggest that there was no difference in our treatments.

7

Wind Pollination in the Humboldt Bay Wallflower We emasculated all of the plants on the ridge on March 25th when it was raining and windy. These plants were more prone to damage during emasculation due to these harsh conditions. This damage may have impacted the level of receptivity of the stigmas and resulted in an underestimation of the number of pollen grains normally recieved by wind. Plants in our hollow treatment group were emasculated on March 26 when weather was clear and of moderate temperatures, thus they were not exposed to the same conditions as the ridge plants had been immediately after emasculation. The weather was constant for the remainder of our study period. Lasioglossum is the only known visitor of Erysimum small enough to pass through the mesh of our cages, but Lasioglossum is skittish and cages could potentially discourage visitation by acting as a physical barrier around the plant. Lasioglossum was commonly observed at the field site on Erysimum, but was never seen inside or near the cages. We did see ants and Elaterid beetles on some of the plants. While these insects could potentially move pollen from anthers to stigmas within a flower, or between flowers on a single plant, it is unlikely that they could move pollen between plants. The integument of these insects is smooth and pollen does not seem to stick to it well. We collected voucher specimens of pollen to aid in identifying the pollen found on stigmas. Voucher pollen was made from fresh anthers and mounted in fuchsin gel immediately after collection. Stigmas remained exposed for up to five days and pollen on the stigmas may have become dehydrated and changed shape while they were exposed. As a result, we may have counted some of the Erysimum pollen erroneously as heterospecific pollen because it didn’t match our Erysimum voucher pollen. While some of the heterospecific pollen could be positively identified, the majority of the heterospecific pollen was not similar to any pollen we looked at, and may have been from Erysimum. In order to positively identify it, we would need to examine the texture and arrangement of pores and furrows on the microspore wall at high magnification and compare this with a comprehensive collection from all plants flowering at the Lanphere preserve during our study. Improvements to this study may be able to demonstrate more definitively the role of wind in the pollination of the Humboldt Bay wallflower. The design of the pollen traps should be improved to effectively document the presence of airborne pollen. Wind

8

Wind Pollination in the Humboldt Bay Wallflower speed should be measured to ascertain that there is a real difference between ridges and hollows. If there is no difference in the wind regimes we need to devise a treatment that controls for differing levels of wind. Pollen identification techniques need to be improved to increase the accuracy of pollen counts. Though ants and beetles were not expected to act as pollinators in this experiment, a better cage could exclude them completely. In order to determine the effectiveness of our cages in excluding Lasioglossum, we could conduct more extensive observations of these bees interacting with the cages. The level of contamination in plants due to error in the emasculation technique was not great enough to explain the high pollen loads observed on the stigmas. The plants were caged immediately after emasculation to prevent autogamy, ensuring wind as the only pollen vector (Gomez and Zamora, 1996). Despite the shortcomings of this study, the fact that we found substantial amounts of Erysimum pollen on our stigmas is a significant finding which cannot be explained by contamination. In addition, our Erysimum pollen counts may be conservative. These points demonstrate that Erysimum can collect airborne pollen and suggests that wind may indeed play a substantial role in the pollination of Erysimum.

Acknowledgements We would like to acknowledge Dr. Michael Mesler for his guidance, funding, and support throughout this project. Andrea Pickart provided site maps and use permits to the Lanphere Christensen Unit of the Humboldt Bay National Wildlife Refuge. Anthony Baker supplied microscope coverslips, engineered a specialized microscope slide storage box, and provided other equipment services. Kristal Watrous, Molly Alles, Justin Garwood, James Sclafani, Kim McFarland, and Andrew Jordan helped carry cages and provided moral support. Literature Cited Culley, T. M; S. G. Weller; A. K. Sakai. 2002. The Evolution of Wind Pollination in Angiosperms. TRENDS in Ecology & Evolution. 17(8): 361-369.

9

Wind Pollination in the Humboldt Bay Wallflower Gomez, J. M. and R. Zamora. 1996. Wind Pollination in High-Mountain Populations of Horathophylla spinosa (Cruciferae). American Journal of Botany. 83(5):580-585. Goodwillie, C. 1999. Wind Pollination and Reproductive Assurance in Linanthus parviflorus (Polemoniaceae), A Self-Incompatible Annual. American Journal of Botany. 86(7): 948-954. Kearns, C. A., and D. W. Inouye. 1993. Techniques for Pollination Biologists. University Press of Colorado, Niwot, CO. Whitehead, D. R. 1968. Wind Pollination in the Angiosperms: Evolutionary and Environmental Considerations. Evolution 23: 28-35.

10