Flowering phenology, pollination and seeding interactions in Garden
Lupine (Lupinus polyphyllus)
Relationer mellan blomningsfenologi, pollinering och frösättning hos blomsterlupin (Lupinus polyphyllus)
Amanda Boström
Faculty of Health, Science and Technology Biology
Bachelor´s thesis, 15 hp Supervisor: Lutz Eckstein Examiner: Larry Greenberg 2020-11-02
Series number: 20:181
Abstract
The spreading of the invasive plant Garden Lupine (Lupinus polyphyllus) has become a matter of national importance in Sweden, due to it posing a threat to native plant and pollinator diversity. The effective attraction of bumblebees (Bombus spp.) as pollinators facilitates the production of large numbers of seeds, which are key to the Garden Lupine’s success. Possible self-pollination could also provide a competitive edge for the plant. The objective of this study was to study the relationships between Garden Lupine color morphs, pollinator attraction and seeding. Inflorescences of three color morphs were studied during the flowering period, and bumblebee behavior was observed on the site.
After seeding, any produced seeds were collected and analyzed, as well as experimentally germinated to provide insight into their viability. A subset of inflorescences of each color morph was prevented access to pollinators, to study potential self-pollination effects. Bumblebees preferred blue flowers over pink, but no difference in pollination between the color morphs was found. Flower color did not affect seed production or seed morphology. Self-pollinated inflorescences produced fewer seeds than those with access to pollinators, but no difference in seed morphology or germinative success between the
pollination methods could be established. The results suggest that seed production and germination are less dependent on pollination than expected. The ability to germinate through self-pollination provides insight into the invasive potential of Garden Lupine, suggesting that further studies are needed to successfully counteract its spread.
Sammanfattning
Den invasiva växten blomsterlupin (Lupinus polyphyllus) har på senare år blivit en nationell angelägenhet i Sverige, där den hotar mångfalden av inhemska växter och pollinatörer.
Blomsterlupinens framgångsrika tilldragning av framförallt humlor (Bombus spp.) som pollinatörer möjliggör det stora antalet frön som den producerar, vilket är nyckeln till dess invasiva etablering.
Eventuell förmåga till självpollinering kan också utgöra en konkurrensfördel. Målet med studien var att utforska relationen mellan blomsterlupinens färgmorfer, pollinering samt fröbildning. Blomställningar av tre färgmorfer studerades under blomningsperioden. Humlornas beteende observerades också under perioden vid lupinlokalen. Efter frösättning samlades alla producerade frön upp och analyserades, varefter ett frögroningsexperiment utfördes för att belysa frönas grobarhet. I ett fältexperiment nekades en delmängd av blomställningarna tillgång till pollinatörer, för att studera eventuell självpollinering och dess effekter. Humlorna föredrog blåa blommor före rosa, men ingen skillnad i pollinering mellan färgmorferna kunde fastställas. Blommornas färg hade ingen effekt på fröproduktion eller -morfologi.
Självpollinerade blomställningar producerade färre frön överlag än de med tillgång till pollinatörer, men ingen skillnad i frömorfologi eller grobarhet mellan pollineringsmetoderna kunde påvisas.
Resultaten antyder att fröproduktion och frögroning hos blomsterlupin är beroende av pollinering i mindre grad än förväntat. Förmågan att gro genom självpollinering belyser blomsterlupinens invasiva potential, och antyder att fler studier behövs för att framgångsrikt motverka dess spridning.
Introduction
The introduction of a non-native plant species into an ecological community changes the structure and processes of that community (Bartomeus, Vilá & Santamaria, 2008). Flowering plants and pollinators are engaged in mutualistic networks of coevolution, which may be key in sustaining diversity in many communities. Invasive plants affect plant-pollinator interactions by potentially attracting native pollinators, with reportedly both positive and negative consequences for the native plants of their new community (Bartomeus, Vilá & Santamaría, 2008; Stout & Tiedeken, 2017). The ability to self-pollinate may also improve the reproductive success of invasive species, by reducing their dependency on native pollinators (Stout & Tiedeken, 2017), and the correlation of self-compatibility and a species’ propensity for colonizing has been discussed extensively (Baker, 1955). While self-fertilization is an advantage in terms of short-term survival, it is less advantageous than sexual reproduction when it comes to long- term adaptation to a new environment (Wright, Kalisz & Slotte, 2013).
The flowering perennial Garden Lupine (Lupinus polyphyllus, Fabaceae) has since its introduction to Europe in the nineteenth century quickly become an invasive species, competing with native species in several countries, including Sweden (Fremstad, 2010). Originating from western North America, it was introduced to and popularized in Sweden as a garden plant, due to its colorful and ornamental
inflorescences (Fremstad, 2010). Garden Lupine flowers from late May to early July in Sweden, producing primarily three different color morphs - blue, which is the most common, pink, and white (Pohtio & Teräs, 1995). The inflorescences can grow to over 50 cm, with flowers growing in whorls around a vertical stem. Each inflorescence can produce hundreds of flowers during its flowering period, and each flower has the potential of producing a seed-containing pod. Due to their height and large leaves, they tend to form a dense canopy, obstructing access to both sunlight and pollinators for other flowering plants of smaller stature (Valtonen, Jantunen & Saarinen, 2006). Inflorescences of Garden Lupine flower from the base upwards, with the inflorescence growing as the flowers mature and become pollen-offering. Members of the genus Lupinus have been shown to also produce seeds through self- pollination, albeit to a lesser degree than through open pollination (Shi, Michaels & Mitchell, 2005).
The flowers of Garden Lupine attract many native pollinators, especially bumblebees (Jennersten, Berg
& Lehman, 1988). This is particularly due to their lack of nectar production, hidden pollen, and complex pollen dispensing mechanism (Mossberg & Cederberg, 2012). The detrimental effects of Garden Lupine on the pollination of native species has been studied extensively, showing that during its flowering period, bumblebees seem to prefer lupines to other native, pollen-producing plants. However, the presence of Garden Lupines may also increase the number of native pollinators sustained by the plant community, with net beneficial effects for native plants outside the lupine flowering period (Jakobsson
& Padrón, 2014). Garden Lupines may also adversely affect specialized pollinators of other plant species indirectly, by outcompeting their host plants (Valtonen, Jantunen & Saarinen, 2006). Due to their long tongues and heavy bodies, bumblebees are especially effective at gathering pollen from lupines,
compared to other insects. Despite the lupine flowers hiding their pollen, bumblebees seem to ignore inflorescences with low amounts of available pollen and can distinguish flowers without any pollen left from those that have available pollen (Goulson, Hawson & Stout, 1998). Bumblebees show a preference towards the color of flower which they are first to encounter on a flight, providing they receive a reward from that flower (Gumbert, 2000). They are known to traverse lupine inflorescences upward along the maturation gradient, circling each open whorl. They manipulate the lupine pollen-dispensing
mechanism by pressing their head against the flower banner and, using their legs and pushing the wing petals down, triggering the piston-like mechanism to release pollen out of the keel onto their stomach (Haynes & Mesler, 1984).
Upon pollination, each inflorescence of Garden Lupine produces seed-containing pods, which release the seeds around the mother plant in late summer (Fremstad, 2010). Variation in seed morphology and seed number has been observed in Garden Lupine (Aniszewski, Kupari & Leinonen, 2001; Shi, Michaels
& Mitchell, 2005), and the effects of seed number and size on germination success has been the subject of previous studies of several grassland plants (Jakobsson & Eriksson, 2000). The effects of seed morphology on the germination success as well as germination date of Garden Lupine has also been studied (Klinger, Eckstein, Horlemann, Otte & Ludewig, 2020). The interaction between different color morphs of the Garden Lupine, their pollination (open and self-sustained), seeding and germination has not been studied before, however.
The objective of the study was to explore the following hypotheses:
1: Different color morphs of the Garden Lupine are equally visited by pollinating bumblebees. Due to previous studies on the preferences of bumblebees and the general dominance of blue lupines, I expected pollinators to show a preference towards blue inflorescences.
2: Different color morphs of the Garden Lupine have similar seed production. As stated above, as blue inflorescences are shown to be most prevalent of the Garden Lupine colors, I expected them to show an advantage in seed production.
3: Seed production of Garden Lupine after self-pollination is not significantly different from seed production after open pollination. Due to its propensity for spreading to new areas, I expected my results to show successful seeding through self-pollination to some extent.
4: Germination success of Garden Lupine after self-pollination is not significantly different from germination success after open pollination. I expected both seeds produced through open and self- pollination to produce viable seeds.
Materials and methods
The study was divided into three phases - studying flowering and pollination in the field, followed by seed analysis and a germination experiment performed in the lab. A site containing all three color morphs of Garden Lupine was chosen for the study. The site, a gravel pit in Närkes Kil, Örebro
municipality (59.388763 N, 15.083714 E), had lupines growing extensively and freely. The subset of the site (50x15m) chosen for the study contained 82% blue, 17% pink, and 1% white inflorescences. Seventy- nine Garden Lupine individuals were chosen for the study and divided into three groups - open
pollination, self-pollination, and a control group. These groups consisted of 49, 15, and 15 individuals respectively, each including all three colors. Individuals in the self-pollination and control groups were enclosed in non-woven fabric to keep pollinators, especially bumblebees, out, to study the potential effects of self-pollination of the Garden Lupine. The fabric enclosing individuals in the control group was shredded but kept on the inflorescences for the entire flowering period, to account for the potential adverse effects of the fabric itself on the plants’ performance, while allowing pollinators access to the flowers. All the lupine individuals in the study were numbered and marked with flower pins, facilitating continuous monitoring. Three temperature loggers (Onset, 2020) were placed out on the site, recording the temperature every fifteen minutes, as temperature fluctuations were assumed to affect both
pollinator behavior as well as the rate at which the plants matured (figure A in appendix A). Inventories of flowering plants were performed on 20200531, 20200610 and 20200627, and the presence of bumblebee species on the site was monitored continuously (table A in appendix B).
The site and the lupine individuals for the study were prepared before any flowers matured and opened.
The site was visited and as the inflorescences continued to mature, they grew and the number of flowers offering pollen to pollinators changed. The number of open flowers was assumed to affect the
propensity for pollinators visiting, which is why both the size of the inflorescence and the proportion of open and pollen-offering whorls of flowers on each inflorescence were monitored. All inflorescences in the open pollination group were monitored from when they started maturing until they either lost their flowers entirely or no longer had any flowers open for pollination left. This part of the study was
designed to investigate hypotheses 1 and 2.
Pollination behavior was also observed at each visit to the site. During each visit, randomly selected individuals out of the open pollination group were observed for 10 minutes each, monitoring any visiting pollinators. Species of bumblebee, duration of their stay on each inflorescence and how many individual flowers they visited were logged. The objective of this was to observe whether bumblebee individuals prefer certain color morphs over others, and whether they seemed to visit other plants than Garden Lupines. During the flowering period, each color of Garden Lupine was observed 42 times, making the total observation time 21 hours. In addition to observing pollination of plant individuals within the study, general bumblebee behavior on the site was also observed during each visit. Randomly selected bumblebee individuals were observed for 11-260 seconds, noting which colors of flowers they chose, durations of their visits to each inflorescence and each flower, as well as their method of
manipulating the lupines’ pollen-dispensing mechanism. Color frequencies of lupine flowers, which were later used for analysis, were estimated by counting inflorescences within a set perimeter of 25x25m on site. The observation of pollination visits to inflorescences included in the study and the observation of bumblebee transitions between color morphs were designed to investigate hypothesis 1.
Once all of the lupines in the study, including those in the self-pollination and control groups, produced seed pods and they were deemed mature, i.e. became dark brown, hard, and dry, they were collected for further analysis. The collected seed pods were allowed to dry at room temperature for 3 days before each individual’s pods were opened, and the contained seeds were counted, measured at their longest dimension, and weighed. The analysis of seeds was designed to investigate hypotheses 2 and 3.
After analysis of the produced seeds, they were stored in a cool environment at 5℃ for 8 weeks, to simulate winter conditions. After their cold treatment, they were placed in petri dishes with moistened filter paper, one for each plant individual. For individuals that produced more than 50 seeds, a subset of 50 were randomly selected for the experiment. The seeds were then kept in a temperature-controlled climate chamber at 20℃ and provided 8h of light per day for 3 weeks to allow for germination. Due to most of the seeds proving impermeable to water, after five days they were scarified through a small incision on the seed coat. During the experiment, any germinated seeds were counted and removed from the petri dish, and the filter paper in the dish was moistened when needed. At the end of the experiment period, any remaining seeds were inspected for viability by assessing the condition of their contents. The germination experiment was designed to investigate hypothesis 4.
For analyzing the effect of pollination method and inflorescence color, and the potential interaction effect thereof, a two-way ANOVA was used. Prior to this, the homogeneity of variance for the data was tested using a Levene’s test. For seed size and the total amount of seeds per inflorescence, the raw values were transformed using natural logarithms to fit the requirements of the two-way ANOVA. When a significant effect could be demonstrated, pairwise tests were done using a post hoc Tukey’s range test.
Due to the experiment site producing very few white inflorescences, the group of white inflorescences with access to pollinators included in the study was smaller than the blue and pink groups. To ensure that the smaller data set of white inflorescences did not distort the results, all statistical tests were also performed excluding the white group. However, as the results were not significantly different, I only report the results of analyses including all three groups here.
The observed color preferences of bumblebees were analyzed using a chi-squared test, testing whether an association between the previously visited flower’s color and the color of the next flower existed, based on the observed frequencies of lupine colors on the site, which provided an expected visitation rate for each color. The strength of the association between the previously visited flower’s color and the next one was measured using a Cramer’s V test. The odds ratio was also calculated to analyze the measure of association between blue and pink flowers and the bumblebees’ transitions between them.
All statistical analyses were performed using IBM SPSS Statistics 25.
Results
Flowering
The inflorescences of the Garden Lupine grew throughout the study period (figure 1). The white inflorescences on the site started maturing their flowers later than the blue and pink inflorescences in the study and had open whorls later than other colors. All colors had most open whorls in the beginning of June (figure 2), following a period of relatively high precipitation (figure B in appendix A).
Figure 1. Mean inflorescence size (mm) during the flowering period of each color of Garden Lupine, error bars: ± 1 SE.
Figure 2. Mean number of open whorls during the flowering period of each color of Garden Lupine, error bars: ± 1 SE.
Pollination
Out of the bumblebee species observed on the site, eight were observed engaging in pollination of Garden Lupine: garden bumblebee (B. hortorum), heath bumblebee (B. jonellus), red-tailed bumblebee (B. lapidarius), white-tailed bumblebee (B. lucorum), common carder bee (B. pascuorum), broken- belted bumblebee (B. soroeensis), short-haired bumblebee (B. subterraneus) and buff-tailed bumblebee (B. terrestris). In general, the larger species like the short-haired bumblebee and the buff-tailed
bumblebee, as well as the queens of many of the other species, were observed manipulating the lupine flowers’ pollen dispensing mechanism using their hindlegs and abdomen. The smaller species like the heath bumblebee, as well as the workers of some of the larger species, were instead observed using vibration pollination, after “diving” into the flowers. In addition, several species of Apis spp. were observed visiting Garden Lupine flowers. Five of the bumblebee species mentioned above, as well as several species of bees, were observed pollinating the inflorescences included in the study (figure 3).
Figure 3. Percentage of visits made to Garden Lupine inflorescences by each bee species during pollination.
The observed visits to the inflorescences in the study lasted, on average, 14.6 seconds, during which a mean of 5.8 flowers were pollinated. Due to the low number of observations of pollinator visits gathered, no further analysis could be made. No pink inflorescences were observed being visited by pollinators during the study period (figure 4).
Figure 4. Mean number of flowers visited on each inflorescence during pollination by bumblebee species’, arranged by color, error bars: ± 1 SE. Error bars are not shown for single observations.
A chi-square test was used for analyzing the preference of inflorescence color by the bumblebees on the site. The analysis was based on transitions between blue and pink inflorescences, as no visits to white inflorescences were observed. The association between which color (blue/pink) the bee had previously visited and the chosen color of the following inflorescence was significant, 2= 45.56, df = 1, p < 0.001.
Cramér's V showed that the effect of the color of the previously visited inflorescence on the next one was moderate,c= 0.429, p < 0.001. The odds of choosing a blue inflorescence after visiting a blue one was 1.815, while the odds of choosing a pink inflorescence after visiting a blue one was 0.123. The odds of a bumblebee choosing a blue inflorescence over a pink one after visiting a blue one was thus 14.786. This seems to indicate that blue inflorescences of Garden Lupine are preferred by bumblebees, rejecting hypothesis 1. However, observations of the inflorescences in the study could not confirm that the blue flowers were more frequently visited than the other colors (figure 4).
Seeding
Out of all of the inflorescences included in the experiment, 31.6% did not produce any seeds, either through all of the flowers dropping off before seeding could occur, or through no viable seeds being produced in the pods that did form. Of the color morphs, 29.4% of blue, 40.6% of pink and 15.4% of white flowers did not produce any seeds. On average, each inflorescence produced 65.23 (± 9.05 SE) seeds. No white inflorescences in the self-pollinated group produced any seeds (figures 5-8).
Figure 5. Mean total number of seeds produced per inflorescence, arranged by color and experiment group, error bars: ± 1 SE. Error bars are not shown for single observations.
Inflorescences with access to pollinators produced significantly more total seeds than those produced through self-pollination (figure 5). Levene’s test showed that the assumption of equal variances for total seeds produced could not be rejected (F6, 46 = 2.27, p = 0.053). Using a two-way ANOVA, neither the interaction of flower color and pollination method (F3, 46 = 0.96, p = 0.42), nor flower color (F2, 46 = 1.78, p
= 0.96), had any demonstrable effect on the total number of seeds produced by inflorescences.
Pollination method had a demonstrable effect on the total number of seeds, however (F2, 46 = 6.52, p = 0.003). Tukey’s post hoc test showed that inflorescences with access to pollinators produced on average 83.23 more seeds than those with restricted access to pollinators (p = 0.001).
Figure 6. Mean number of seeds per pod, arranged by color and experiment group, error bars: ± 1 SE. Error bars are not shown for single observations.
Inflorescences with access to pollinators produced significantly more seeds per pod than those produced through self-pollination (figure 6). Levene’s test showed that the assumption of equal variances for mean seeds in a pod could not be rejected (F6, 70 = 0.593, p = 0.735). Using a two-way ANOVA, no interaction effect on the mean number of seeds in each pod could be demonstrated between flower color and pollination method (F4, 70 = 1.56, p = 0.20). Flower color did not have any significant effect on the mean number of seeds in a pod (F2, 70 = 0.99, p = 0.38). The pollination method did have a significant effect on the number of seeds in a pod, however (F2, 70 = 4.89, p = 0.01). Tukey’s post hoc test showed that flowers with access to pollinators produced on average 2.10 more seeds per pod than those without access to pollinators (p = 0.003).
Figure 7. Mean seed size (mm) per inflorescence, arranged by color and experiment group, error bars: ± 1 SE.
Error bars are not shown for single observations.
No significant difference in seed size could be shown between the different means of pollination or inflorescence color (figure 7). Levene’s test showed that the assumption of equal variances for mean seed sizes could not be rejected (F6, 46 = 0.92, p = 0.49). Using a two-way ANOVA, an interaction between color and pollination method could not be demonstrated for seed size (mm) (F3, 46 = 0.33, p = 0.84). No main effect of flower color (F2, 46 = 0.04, p = 0.97) or pollination method (F2, 46 = 0.59, p = 0.56) on seed size could be demonstrated either.
No significant difference in seed mass could be demonstrated between the different means of
pollination or inflorescence color. Levene’s test showed that the assumption of equal variances for mean seed weight could not be rejected (F6, 70 = 2.20 p = 0.053). No interaction effect between flower color and pollination method on mean seed weight (g) could be demonstrated (F4, 70 = 0.52, p = 0.72). No main effect of flower color (F2, 70 = 2.17, p = 0.12) or pollination method (F2, 70 = 2.85, p = 0.07) on mean seed weight could be demonstrated either.
The results above indicate that the color of the inflorescence does not have a significant effect on seed production, and thus hypothesis 2 can not be rejected. Whether the seeds are produced through open pollination or self-pollination does seem to have a significant effect, however, at least on the total amount of seeds produced and the number of seeds per pod. Hypothesis 3 can thus be rejected.
Germination
Out of the individuals in the germination experiment, 20.4% germinated within the allotted time period, 79.6% still had seeds left that had potential to germinate, while 37.0% had seeds left that were deemed without potential to germinate. Levene’s test showed that the assumption of equal variances for mean percentages of germinated seeds could not be rejected (F6, 46 = 1.66, p = 0.152). Using a two-way ANOVA, no interaction effect on the mean percentage of germinated seeds could be demonstrated between flower color and pollination method (F3, 46 = 0.312, p = 0.817). No main effect of flower color (F2, 46 = 0.284, p = 0.754) or pollination method (F2, 46 = 0.026, p = 0.975) on the percentage of germinated seeds could be demonstrated either. Due to the results stated above, hypothesis 4 could not be rejected.
Discussion
The flowers of Garden Lupine were visited by eight different species of bumblebees, as well as several species of bees. Visits by bumblebees to each inflorescence lasted on average 14.6 seconds, during which on average 5.8 flowers were visited. Bumblebees showed a preference for blue flowers over pink ones, during transition from one inflorescence to the next. Almost a third (31.6%) of inflorescences, including all the white inflorescences in the self-pollinated group, did not produce any seeds. Pollination method had a significant effect on the total number of seeds produced per inflorescence as well as the number of seeds produced per pod, where those with access to pollinators produced on average 83.23 more seeds per inflorescence, and 2.1 more seeds per pod, than those produced through self-pollination. A fifth (20.4%) of the seeds germinated during the experiment period, demonstrating no significant difference in germinative success between the pollination groups. No significant effect of flower color on seeding or germination could be demonstrated, however.
The site contained a wide array of bumblebee species, and almost all of them were observed visiting Garden Lupine flowers over any other flowers in the area. They did not seem to have any difficulty distinguishing between spent flowers and those that were pollen-offering, confirming previous studies on bumblebee behavior (Goulson, Hawson, Stout, 1998). Both bumblebees and honeybees can also distinguish flowers which they themselves, their siblings, or other bees, even of other species have recently visited, and avoid those (Goulson, Hawson, Stout, 1998). Bumblebees of different species and sizes had different approaches to manipulating the lupine pollen-dispensing mechanism, but no
discernable variations in visitation efficacy, however. Studying lupine-pollinator interactions in frontier areas of the Garden Lupine’s spread northward might show more about if and how novel bumblebee species adapt to pollinating the Garden Lupine’s complex flowers. The effective recruitment of such a wide variety of pollinator species is compelling, considering the evidence of Garden Lupine’s ability to sustain large populations of pollinators, with effects lasting after their flowering period is over,
benefiting native flowering plants (Jakobsson & Padrón, 2014).
The pollination frequency of different color morphs of Garden Lupine when observing specific
inflorescences did not differ, according to my findings. On the other hand, when observing bumblebee behavior, a preference for blue flowers over pink flowers could be demonstrated. This confirms previous studies, showing that bumblebees prefer the color of the first rewarding flower they encounter on their foraging trip (Pohtio & Teräs, 1995). As the site in the study contained 82% blue lupines, the chance of any bumblebee encountering a blue rewarding flower on their flight is high. After associating a flower color with a reward, bumblebees also prefer other colors with similar perceptual values, which can differ from the colors perceived by human eyes (Gumbert, 2000). To some extent, almost all bumblebee and honeybee species also show an innate preference towards colors within 400-420 nm on the color spectrum, which corresponds to blue and purple shades, further confirming the findings in my study and previous studies that show a preference towards blue lupine flowers. Bumblebees also prefer different colors depending on the season and weather (Pohtio & Teräs, 1995). Color preferences in bumblebees and how it affects pollination of Garden Lupine is worth exploring further.
While the pollinators preferred blue flowers over pink, flower color did not seem to affect seed production or seed size, contrary to my expectations. Despite this, blue flowers at the site and in the general population are much more common than white or pink (Pohtio & Teräs, 1995). As seeds produced through open pollination were significantly more numerous than those produced through self-pollination, I would expect that flowers more effectively pollinated would dominate the population.
My findings showed that while a preference between colors exists, pollinators do not visit blue flowers more often than any other color. On the other hand, pollinator visits to pink flowers in the study were never observed. The means of pollination had a significant effect on the number of seeds produced, but not on seed mass or size. Previous studies have shown no relation between seed size and seed number in Garden Lupine but have found a relationship between seed size and germinative success, as well as
seedling size. Having a seed size that is relatively stable despite environmental factors like pollinators could prove key to the Garden Lupines invasive success (Sõber & Ramula, 2013).
The Garden Lupine evidently has the ability to produce seeds through self-pollination, as shown in the results. While seed production through self-pollination is less effective than through open pollination, the ability to self-pollinate at all could provide a key to understanding the invasive success of the Garden Lupine. Self-reliance in producing seeds reduces dependency on native pollinators when introduced to new frontiers (Stout & Tiedeken, 2017), and could potentially enable Garden Lupine to colonize areas without large populations of pollinators. In such a scenario, a newly established Garden Lupine population, initially sustained through self-pollination, could potentially recruit and sustain more pollinators in an area where they were previously few in numbers (Jakobsson & Padrón, 2014).
The germination experiment showed that seeds produced through self-pollination have a similar germinative success as seeds produced through open pollination, which further establishes the possibility of Garden Lupines potentially utilizing the ability to self-sustain when needed.
To further investigate the potential of self-pollination and the co-existence of Garden Lupines and bumblebees, a similar study would benefit from upscaling to involve larger sample sizes in all three groups - open pollination, self-pollination, and control. On a smaller scale such as this study was performed, any outliers in the data are assigned proportionally larger significance than they probably should receive. Especially the lack of pollinator visits to pink flowers, and the fact that nearly all of the white flowers on the site were a part of the study due to them being so few overall, are aspects that could be very different provided a larger scale. Correlating pollination method and flower color to seed
morphology, as well as germinative success over an extended period might also produce interesting results. Seed color and hardness, for example, are good indicators of germinative success in Garden Lupine (Klinger, Eckstein, Horlemann, Otte & Ludewig, 2020), and potential areas of study for the future.
The site itself is in many ways very different to many of the typical sites for Garden Lupine studies. The area of the gravel pit where lupines are growing has been left untouched for years, according to the landowner. Frequently mowed roadside populations could produce different results compared to the site used here, especially considering recent studies showing the importance of mowing dates on the outcomes of Garden Lupine germination and growth (Klinger, Eckstein, Horlemann, Otte & Ludewig, 2020). The spreading of Garden Lupines has led to local loss of native and already threatened flowering plants that have found refuge from agricultural practices in road verges. Preserving these flowering plants has become a matter of great importance for many national institutions in Sweden. One of the key ways of combating their spreading is through frequent mowing of road verges, conducted in Sweden primarily by Trafikverket. Mowing can also result in the facilitation of seed dispersal, however, if
performed at the wrong time during seed development, which is why the relationship between seed morphology and seed germination is key to conservation (Wissmann, Norlin & Lennartsson, 2015).
In conclusion, insight into the interactions of pollination and seeding can provide clues as to how best to counteract the spreading of invasive species that outcompete native populations of endangered flowering plants. Preserving plant diversity along road verges in Sweden has in recent years been elevated to a matter of national importance, with attempts made to reach out to the public about invasive species, the threats they present, and how everyone can contribute in counteracting their spread. A national day for combating the Garden Lupine has been declared (Fältbiologerna, 2020).
Finland has officially added the Garden Lupine to the national list of invasive species, prohibiting selling and growing them (Vieraslajit, 2019). The general advice given to the public is to cut the lupines before they produce seeds, and to dispose of their seeds in a way that prevents them from germinating (Naturvårdsverket, 2020). I believe that future studies into the relationship between pollination, seeding and germination of Garden Lupine will yield more successful conservation efforts.
References
Aniszewski, T., Kupari, M. H., Leinonen, A. K. (2001). Seed number, seed size and seed diversity in Washington Lupin (Lupinus polyphyllus Lindl.). Annals of Botany, 87(1), 77-82.
https://doi.org/10.1006/anbo.2000.1300
Baker, H., G. (2005). Self-compatibility and establishment after ‘long-distance’ dispersal. Evolution, 9(3), 347-349. https://doi.org/10.2307/2405656
Bartomeus, I., Vilá, M., Santamaría, L. (2008), Contrasting effects of invasive plants in plant-pollinator networks. Oecologia, 155(1), 761-770. https://doi.org/10.1007/s00442-007-0946-1
Fremstad, E. (2010). NOBANIS, Invasive Alien Species Fact Sheet, Lupinus polyphyllus. Online
Database of the European Network on Invasive Alien Species – NOBANIS. Retrieved 2020-05-18 from https://www.nobanis.org/globalassets/speciesinfo/l/lupinus-polyphyllus/lupinus-polyphyllus.pdf
Fältbiologerna. (2020). Stora lupinbekämpardagen - för biologisk mångfald. Retrieved 2020-10-20, from https://www.faltbiologerna.se/evenemang/stora-lupinbekampardagen-for-biologisk-
mangfald/#:~:text=Den%20sj%C3%A4tte%20juni%2C%20p%C3%A5%20Sveriges,landet%20f%C3%B6r%2 0att%20bek%C3%A4mpa%20lupinerna!
Goulson, D., Hawson, S. A., Stout, J. C. (1998). Foraging bumblebees avoid flowers already visited by conspecifics or by other bumblebee species. Animal behaviour, 55(1), 199-206.
https://doi.org/10.1006/anbe.1997.0570
Gumbert, A. (2000). Color choices by bumble bees (Bombus terrestris): innate preferences and generalization after learning. Behavioral Ecology and Sociobiology, 48(1), 36-43.
https://doi.org/10.1007/s002650000213
Haynes, J., Mesler, M. (1984). Pollen foraging by bumblebees: Foraging patterns and efficiency on Lupinus polyphyllus. Oecologia, 61(2), 249-253. https://doi.org/10.1007/bf00396768
Jakobsson, A., Eriksson, O. (2000). A comparative study of seed number, seed size, seedling size and recruitment in grassland plants. OIKOS, 88(3), 494-502. https://doi.org/10.1034/j.1600-
0706.2000.880304.x
Jakobsson, A., Padrón, B. (2014). Does the invasive Lupinus polyphyllus increase pollinator visitation to a native herb through effects on pollinator population sizes? Oecologia, 174(1), 217-226.
https://doi.org/10.1007/s00442-013-2756-y
Jennersten, O., Berg, L., Lehman, C. (1988). Phenological differences in pollinator visitation, pollen deposition and seed set in the sticky catchfly, Viscaria vulgaris. Journal of Ecology, 76(4), 1111-1132.
https://doi.org/10.2307/2260638
Klinger, Y. P., Eckstein, R. L., Horlemann, D., Otte, A., Ludewig, K. (2020). Germination of the invasive legume Lupinus polyphyllus depends on cutting date and seed morphology. Neobiota 60(1), 79-95.
https://doi.org/10.3897/neobiota.60.56117
Mossberg, B., Cederberg, B. (2012). Humlor i Sverige (1st ed.). Bonnier Fakta.
Naturvårdsverket. (2020). Blomsterlupin (Lupinus polyphyllus). Retrieved 2020-10-14, from https://www.naturvardsverket.se/Sa-mar-miljon/Vaxter-och-djur/Frammande-arter/Invasiva- frammande-arter/Arter-som-inte-ar-EU-reglerade/Blomsterlupin-/
Onset. (2020). HOBO Pendant® MX Water Temperature Data Logger. Retrieved 2020-09-12, from https://www.onsetcomp.com/products/data-loggers/mx2201/
Pohtio, I., Teräs, I. (1995). Bumblebee visits to different colour morphs of the Washington lupine, Lupinus polyphyllus. Entomologica Fennica, 6(2-3), 139-151. https://doi.org/10.33338/ef.83851 Shi, X. J., Michaels, H. J., Mitchell, R. J. (2005). Effects of self-pollination and maternal resources on reproduction and offspring performance in the wild lupine, Lupinus perennis (Fabaceae). Sexual Plant Reproduction, 18(1), 55-64. https://doi.org/10.1007/s00497-005-0250-3
SMHI, Sveriges meteorologiska och hydrologiska institut. (2020a). Nederbörd, Kilsbergen Suttarboda A. Retrieved 2020-08-24, from https://www.smhi.se/data/meteorologi/ladda-ner-meteorologiska- observationer/#param=precipitation24HourSum,stations=all,stationid=94190
SMHI, Sveriges meteorologiska och hydrologiska institut. (2020b). Vind, Kilsbergen Suttarboda A.
Retrieved 2020-08-24, from https://www.smhi.se/data/meteorologi/ladda-ner-meteorologiska- observationer/#param=wind,stations=all,stationid=94190
Sõber, V., Ramula, S. (2013). Seed number and environmental conditions do not explain seed size variability for the invasive herb Lupinus polyphyllus. Plant Ecology, 214(1), 883-892.
https://doi.org/10.1007/s11258-013-0216-8
Stout, J. C., Tiedeken, E. J. (2017). Direct interactions between invasive plants and native pollinators:
evidence, impacts and approaches. Functional Ecology, 31(1), 38-46. https://doi.org/10.1111/1365- 2435.12751
Valtonen, A., Jantunen, J., Saarinen, K. (2006). Flora and lepidoptera fauna adversely affected by invasive Lupinus polyphyllus along road verges. Biological conservation, 133(3), 389-396.
https://doi.org/10.1016/j.biocon.2006.06.015
Vieraslajit. (2019). Vieraslajiasetus 704/2019. Retrieved 2020-10-14, from https://vieraslajit.fi/fi/content/vieraslajiasetus-7042019
Wissmann, J., Norlin, K., Lennartsson, T. (2015). Invasiva arter i infrastruktur. Centrum för biologisk mångfald. Retrieved 2020-10-13, from
https://www.slu.se/globalassets/ew/org/centrb/cbm/dokument/publikationer-cbm/cbm- skriftserie/invasiva-arter-i-infrastruktur.pdf
Wright, S. I., Kalisz, S., Slotte, T. (2013). Evolutionary consequences of self-fertilization in plants.
Proceedings, Biological Sciences, 280(1760), 20130133. https://doi.org/10.1098/rspb.2013.0133
Appendices
Appendix A - Weather data
Figure A. Mean temperature (°C) during flowering and seeding of Garden Lupines on site. Measured locally using HOBO Temperature Data Loggers.
Figure B. Mean precipitation (mm) during flowering and seeding of Garden Lupines on site. Measured at Kilsbergen-Suttarboda A measuring station (SMHI, 2020a).
Figure C. Mean wind speed (m/s) during flowering and seeding of Garden Lupines on site. Measured at Kilsbergen-Suttarboda A measuring station (SMHI, 2020b).
Appendix B - The site
Figure D. The site with extensive Garden Lupine growth chosen for the study.
Table A. Observed species, bumblebees (Bombus spp.) and flowering plants, on the site of the study.
* flowered simultaneously with Garden Lupine
** flowered after Garden Lupine
Scientific name Swedish name Common name Bombus terrestris Mörk jordhumla Buff-tailed bumblebee Bombus lucorum Ljus jordhumla White-tailed bumblebee Bombus jonellus Ljunghumla Heath bumblebee
Bombus lapidarius Stenhumla Red-tailed bumblee Bombus subterraneus Vallhumla Short-haired bumblebee Bombus soroeensis Blåklockshumla Broken-belted bumblebee Bombus pascuorum Åkerhumla Common carder bee Bombus hortorum Trädgårdshumla Garden bumblebee Bombus pratorum Ängshumla Early bumblebee Bombus ruderarius Gräshumla Red-shanked carder bee Lotus corniculatus* Käringtand Birds-foot trefoil Ranunculus acris* Smörblomma Meadow buttercup
Fragaria vesca* Smultron Wild strawberry
Taraxacum spp.* Maskrosor Dandelions
Anthriscus sylvestris* Hundkäx Cow parsley Silene viscaria* Tjärblomster Sticky catchfly Campanula patula* Ängsklocka Spreading bellflower Stellaria graminea* Grässtjärnblomma Common starwort Melampyrum nemorosum* Natt och dag
Arabidopsis suecica* Grustrav
Potentilla argentea* Femfingerört Hoary cinquefoil Veronica chamaedrys* Teveronika Germander speedwell
Rubus idaeus* Hallon Raspberry
Aquilegia vulgaris* Akleja European columbine Chamaenerion angustifolium** Mjölkört Fireweed
Leucanthemum vulgare** Prästkrage Oxeye daisy
Galium album** Stormåra White bedstraw
Trifolium pratense** Rödklöver Red clover
Campanula persicifolia** Stor blåklocka Peach-leaved bellflower Achillea millefolium** Röllika Common yarrow Sonchus arvensis** Åkermolke Field milk thistle
