The Pollination Biology of Haskap (Lonicera Caerulea L.: Caprifoliaceae): Floral Traits

The pollination biology of Haskap (Lonicera caerulea L.: Caprifoliaceae): floral traits

and pollinator performance of a new Saskatchewan fruit crop.

A Thesis

Submitted to the Faculty of Graduate Studies and Research

In Partial Fulfillment of the Requirements

For the Degree of

Master of Science

Biology

University of Regina

Sandra Danae Frier

Regina, Saskatchewan

February, 2016

UNIVERSITY OF REGINA

FACULTY OF GRADUATE STUDIES AND RESEARCH

SUPERVISORY AND EXAMINING COMMITTEE

Sandra Danae Frier, candidate for the degree of Master of Science in Biology, has presented a thesis titled, The pollination biology of Haskap (Lonicera caerulea L.: Caprifoliaceae): floral traits and pollinator performance of a new Saskatchewan fruit crop., in an oral examination held on December 16, 2015. The following committee members have found the thesis acceptable in form and content, and that the candidate demonstrated satisfactory knowledge of the subject material.

External Examiner: Dr. Art Davis, University of Saskatchewan

Co-Supervisor: Dr. Christopher Somers, Department of Biology

Co-Supervisor: Dr. Cory Sheffield, Department of Biology

Committee Member: Dr. Harold Weger, Department of Biology

Chair of Defense: Dr. Rozzet Jurdi, Department of Sociology and Social Studies

ABSTRACT

Haskap (Lonicera caerulea L.) is an early flowering fruiting shrub native to northern regions of Europe, Asia and North America. Commercial growth began in

Europe and Asia over a century ago, but it has only recently been cultivated in North

America. It is self-incompatible, and requires insect pollinators in order to set fruit; however, very little is currently known about its pollination biology, including associated pollinators and important floral characteristics. My research compared the pollinating performance of three groups of bees on Haskap: commercial Apis mellifera and Osmia lignaria, and wild Bombus spp. I found that Bombus queens had the highest levels of single visit pollen deposition (SVD), the highest visitation rate, and were observed to be active even during cooler temperatures experienced during Haskap flowering. Apis mellifera workers had the lowest levels of SVD, spent nearly three times as long per flower as Bombus spp., and were not active during cooler temperatures; however, as they were present in much higher densities than Bombus, they may be effective in supplementing pollination in pollinator-scarce environments. Osmia lignaria females were found to have potentially high levels of SVD and intermediate visitation rates, however, pollen load analysis suggests they prefer alternative forage and are not frequent visitors to Haskap in field conditions. In addition to pollinator performance, I describe anthesis length, nectar and pollen dynamics, and stigma receptivity in Haskap flowers. I observed that un-pollinated flowers bloomed for 3-4 days, but pollination triggered early senescence. Nectar was present at the onset of anthesis and did not increase significantly after 16 hours, and any nectar removed during anthesis was replaced. There was no evidence of nectar resorption. Pollen release began immediately after the flower opened

i and continued over the first day, and the stigma showed evidence of receptivity while still in the bud stage. These results suggest that Haskap flowers exhibit a generalist pollination strategy that optimizes pollination opportunities in pollinator scarce environments. Observations that some flowers open in the evening (and possibly overnight) suggest that nocturnal pollinators such as moths may be important for the crop, in addition to known diurnal pollinators.

ACKNOWLEDGEMENTS

This thesis would not have been possible without a long list of people, all of which I am incredibly thankful to. First and foremost, a big thank you to my supervisors

Dr. Cory S. Sheffield and Dr. Christopher M. Somers, for giving me guidance and many opportunities to grow as an individual and a scientist. I would also like to thank Dr.

Harold Weger for serving on my committee and his valuable input to my project.

My field assistants, Branden Henry, Graham Rothwell, Laura Messett, and

Shelby Rosvold, were integral to the success of this project, and I thank them for their hard work and patience. I am indebted as well to Hamish Graham for allowing us to work in his orchard, as well as for his gracious hospitality and professional expertise. I also would like to thank the rest of the members of Haskap Canada and other growers within the province for their continued support and interest in my work.

I would like to thank Dr. Bob Bors for his advice and expertise, and to

Bayartulga Lkhagvasuren for giving me access to greenhouse plants, without which much of this work would not have been possible.

Lastly, I give thanks to the rest of the members of the Somers lab, and especially

Leanne Heisler, for both moral support and professional guidance, without whom it would have been a lonely road.

This work was made possible by funding from the Government of Saskatchewan

Department of Agriculture through the Agriculture Development Fund (ADF) and

Canadian Agriculture Adaptation Program (CAAP).

iii

POST DEFENSE ACKNOWLEDGEMENTS

Dr. Arthur R. Davis graciously agreed to act as my external examiner. His expert and thorough review has significantly improved the quality of the final version of this thesis.

DEDICATION

To my grandpa, Steve, who gave me an enduring love for the outdoors and an appreciation for all living things.

TABLE OF CONTENTS

Abstract ...... i

Acknowledgements ...... iii

Post Defense Acknowledgements ...... iv

Dedication ...... v

Table of Contents ...... vi

List of Tables ...... viii

List of Figures ...... ix

List of Appendices ...... xi

Chapter 1: General Introduction ...... 1 1.1 Background ...... 2 1.2 Floral biology of Haskap ...... 3 1.3 Assessing pollinator performance ...... 4 1.4 Study objectives ...... 7 References ...... 8

Chapter 2: Comparing the performance of native and managed pollinators of Haskap (Lonicera caerulea: Caprifoliaceae), an emerging fruit crop ...... 11 1. Introduction ...... 12 2. Methods ...... 15

2.1 Study site ...... 15

2.2 Single visit deposition ...... 16 2.3 Visit symmetry and duration ...... 17

2.4 Pollinator density ...... 17 2.5 Pollen load composition of Osmia lignaria ...... 18 2.6 Statistical analysis ...... 18 3. Results ...... 20

3.1 Single visit deposition ...... 20

3.2 Visit symmetry and duration ...... 20

3.3 Pollinator density ...... 21 3.4 Pollen load composition of Osmia lignaria ...... 21 4. Discussion ...... 22

5. Conclusions ...... 26 References ...... 27

Chapter 3: Floral longevity, nectar production and pollen release in Haskap (Lonicera caerulea L.: Caprifoliaceae) ...... 37 1. Introduction ...... 38 2. Methods ...... 40 2.1 Study site and plants ...... 40 2.2 Length of anthesis ...... 41 2.3 Nectar, pollen and stigma dynamics ...... 42 3. Results ...... 44 3.1 Length of anthesis ...... 44 3.2 Nectar, pollen and stigma dynamics ...... 44 4. Discussion ...... 45 5. Conclusion ...... 48 References ...... 49

Chapter 4: General Discussion ...... 58 4.1 Synthesis ...... 59 4.2 Conclusions and recommendations ...... 61 4.3 Suggestions for further research ...... 64 References ...... 66 APPENDIX A – Haskap floral traits ...... 71 APPENDIX B – Pollination requirements for fruit set in Haskap ...... 72 B.1 Methods ...... 72 B.2 Results ...... 74

vii

LIST OF TABLES

Chapter 2

Table 1. The results of a negative binomial GLM comparing the number of pollen grains found on the stigmas of Haskap (Lonicera caerulea L.) flowers after a single visit by Apis mellifera L., Bombus spp. or Osmia lignaria Say to an unvisited control...... 32

Table 2. The results of a negative binomial GLM regression on the response of Apis mellifera L. and Bombus spp. densities to the variables time, temperature, wind, cloud cover and density of the other taxon...... 36

Appendix A

Table A.1. Observed floral visitors to Haskap (Lonicera caerulea L.) in an orchard near Birch Hills, Saskatchewan in May, 2013. Individuals were identified through several 30 minute surveys conducted over a three day period, as well as through opportunistic observations and capture events. Bombus spp. were identified on the wing whenever possible. Small solitary bees as well as unidentified Bombus spp. were collected using a sweep net and were stored in ethanol to be identified at a later time. Voucher specimens from this study are stored in the collection at the Royal Saskatchewan Museum, Regina, Saskatchewan, Canada...... 71

Appendix B

Table B.1 The results of generalized linear models comparing length, width, weight and seed number in Haskap (Lonicera caerulea L.) fruit that had either one or both flowers in the inflorescence pollinated...... 76

viii

LIST OF FIGURES

Chapter 2

Figure 1. Haskap (Lonicera caerulea L.) flowers and fruits. (a) A typical Haskap inflorescence. The two perfect flowers each have five petals, five stamens and one stigma; the ovaries of both are enclosed by bracteoles which are fused into a cupule and are subtended by two bracts. The ovaries and bracteoles develop into a single compound fruit. (b) Normal compound fruit produced by Haskap. A single fruit results from each two-flowered inflorescence. (c) Haskap fruit in which the bracteoles have not fused around the ovaries, showing the two distinct berries which make up the fruit (Photo credit: Logie Cassells, LaHave Natural Farms, Nova Scotia, Canada)...... 31

Figure 2. Single visit pollen deposition (SVD) of pollen on Haskap (Lonicera caerulea L.) stigmas by three taxa of bees. SVD is derived from the predicted estimates of a generalized linear model, and represents the total amount of pollen found on visited stigmas minus the pollen on unvisited controls, the result of internal self-pollination. The dotted line represents the minimum pollination requirements (i.e., 24 grains deposited) for each flower to ensure complete fertilization of all the ovules (see Appendix B, Supplementary Material)...... 33

Figure 3. Mean (± SE) densities of Bombus spp. and Apis mellifera L. in a Haskap (Lonicera caerulea L.) orchard at each hour from dawn until dusk during flowering. The data are presented as number of individuals per 400 open flowers...... 34

Figure 4. The relationship between density of foragers and air temperature (˚C) for Apis mellifera and Bombus spp. across 99 surveys performed from 24-May to 10-June-2014 in a 30ha organic Haskap orchard. Surveys were performed by scan sampling, and a total of 400 flowers were observed per survey...... 35

Chapter 3

Figure 1. A typical Haskap inflorescence. The downward facing flowers are pale yellow and each has five petals, five stamens and one stigma; the ovaries of both flowers are enclosed by the bracteoles, which have fused into a cupule. The ovaries and bracteoles develop into the Haskap ‘berry’, actually a multiple fruit...... 53

Figure 2. The length of anthesis (h), from opening of the bud to abscission of the corolla, for un-pollinated and pollinated Haskap inflorescences that had the anthers removed or left intact. Vertical bars represent the 95% confidence interval. Bars with different letters are significantly different at P < 0.05 according to pairwise comparisons

ix using a Mann-Whitney U test with p-values adjusted for multiple comparisons using the Bonferroni correction. Sample size is reported in brackets beneath each bar...... 54

Figure 3. The nectar volume (µL) (left y-axis) and the number of open anthers (right y- Axis) in Haskap (Lonicera caerulea) inflorescences per time interval (x-axis), starting the first morning the flowers were open until the end of flowering (i.e., inflorescences that opened after 9am were begun sampling the following morning). Vertical bars represent the 95% confidence interval. Plots with different letters are significantly different at P < 0.05 according to pairwise comparisons using a Mann-Whitney U test with p-values adjusted for multiple comparisons using the Bonferroni correction. Sample size is indicated in brackets under each treatment...... 55

Figure 4. The amount of nectar (µL) (left y-axis) and the number of open anthers (right y-axis) in Haskap (Lonicera caerulea) inflorescences per 8 hour interval after the onset of anthesis (x-axis). Vertical bars represent the 95% confidence interval. Plots with different letters are significantly different at P < 0.05 according to pairwise comparisons using a Mann-Whitney U test with p-values adjusted for multiple comparisons using the Bonferroni correction. Sample size is indicated in brackets under each treatment...... 56

Figure 5. The amount of nectar (µL) in Haskap (Lonicera caerulea) inflorescences that had nectar removed once per day, up to 3 consecutive days. Vertical bars represent the 95% confidence interval. Plots with different letters are significantly different at P < 0.05 according to pairwise comparisons using a Mann-Whitney U test with p-values adjusted for multiple comparisons using the Bonferroni correction. Sample size is indicated in brackets under each treatment...... 57

Appendix B

Figure B.1 The (a) length (mm), (b) width (mm), (c) weight (g) and (d) number of seeds of Haskap (Lonicera caerulea L.) berries from two cultivars, ‘Indigo Gem’ and ‘Tundra’, in which one (‘Indigo Gem’, n=9; ‘Tundra’, n=9) or both (‘Indigo Gem’, n=9; ‘Tundra’, n=8) of the flowers in the inflorescence were pollinated. Each sample represents the average measurements from 1-5 berries per shrub, and measurements were only taken from berries that were fully ripe at the time of harvest. Lower, middle and upper quartiles are presented by the boxes and maximum and minimum values by the vertical bars...... 75

Figure B.2 Weight (g) vs seed number in Haskap fruit (r=0.73, t=11.05, df=106, p< 0.001)...... 77

LIST OF APPENDICES

APPENDIX A – Haskap floral traits ...... 71

APPENDIX B – Pollination requirements for fruit set in Haskap ...... 72

CHAPTER 1: General Introduction

1.1 Background

The genus Lonicera (Caprifoliaceae) is a group of temperature fruiting shrubs found throughout temperate areas of the world, with the highest diversity occurring in

Central and Eastern Asia. There are around 180 described species, and whereas the majority bear fruits that are inedible to humans, a notable exception is Lonicera caerulea

L. (Miller et al., 2011; Rehder, 1903), commonly known as blue honeysuckle, honeyberry, and Haskap (Hummer et al., 2012). Various subspecies of L. caerulea grow native to northern areas of North America, Europe, and Asia, and the fruits were an important part of the diet of hunter-gatherers in Russia and Hokkaido, Japan, where people have long recognized their nutritional and medicinal value (Hummer et al., 2012;

Thompson, 2006).

Domestication of L. caerulea as a crop began in Russia in the early 1900’s, and intensified in the 1950’s, but until recently it has been largely unknown within North

America (Thompson, 2006). A breeding program at Oregon State University in the late

1990’s began raising interest in this new crop (Bors, 2008); today, one of the largest breeding programs is located at the University of Saskatchewan, in Saskatoon,

Saskatchewan, Canada, which uses germplasm from Russian, Japanese, and Canadian varieties of L. caerulea (Bors et al., 2012). Cultivars developed and released from this program use the name ‘Haskap’ to distinguish them from other cultivars, and to pay homage to the Ainu people from Hokkaido, Japan, who traditionally harvested these berries (Haskap, also spelled Haskappu, Hascap, and Hascup, is the Ainu name used for these berries) (Thompson, 2006).

Recently, the cultivated Haskap industry has been growing in popularity within

North America, largely due to the perceived health benefits of the fruits, which are high in vitamins and anti-oxidants (Chaovanalikit et al., 2004; Raudsepp et al., 2013; Rop et al., 2011; Rupasinghe et al., 2012; Svarcova et al., 2007). The plant is very well suited for cultivation within northern areas of Canada, as it is extremely winter-hardy and is able to withstand temperatures as low as -45˚C during winter dormancy. Haskap is believed to be resistant to most pests and diseases, which makes it a suitable crop for organic management strategies. It flowers very early in the year, before any other fruit crops in Canada, and can be harvested as early as June (hand–picked or mechanically harvested, depending on the cultivar), making it a good companion to other fruit crops.

This early flowering characteristic, alongside its adaptations to cold, make Haskap ill- suited for cultivation in southern areas, as warming periods during the winter could cause it to emerge from dormancy too early (Bors et al., 2012).

1.2 Floral biology of Haskap

Haskap inflorescences have a unique structure that is not seen in any other

Lonicera species (Rehder, 1909); the flowers grow in two-flowered cymes, which is typical of Lonicera, but the bracteoles have fused around the ovaries of both flowers.

The fused bracteoles and ovaries develop into the fruit, meaning that a Haskap “berry” is really a multiple fruit. The flowers themselves are perfect, with 5 anthers and a single stigma. The corolla is pale yellow, about 2.5cm long and tubular in shape (Hummer et al., 2012; Rehder, 1909). Nectar is produced in abundance by nectaries at the bottom of the corolla, and the numerous rougly-oblate pollen grains are ~50µm in diameter (Bożek

3 and Wieniarska, 2006; Bożek, 2007). The flowers are well adapted to spring conditions at high latitudes, and can survive temperatures of at least -7˚C (Bors, 2008; Hummer et al., 2012; Plekhanova, 2000). Despite blooming at a time of year when insects may be scarce, Haskap is essentially self incompatible and requires insects for cross pollination

(Bors et al., 2012; Hummer et al., 2012); Apis mellifera L. (honey bees), Bombus spp.

(bumble bees), and solitary bees have been recorded as visitors to the plant (Bożek,

2012).

Most research on Haskap to date has focused on the fruit; how to make them bigger, better tasting, or easier to harvest (Bors et al., 2012). Very little is known about floral characteristics or pollinators of this crop, and existing studies have been entirely focused in Europe and Asia (Bożek and Wieniarska, 2006; Bożek, 2007, 2012;

Thompson, 2006). Currently, nothing is known about floral visitors to Haskap in North

America, and important aspects of anthesis such as nectar dynamics have not been assessed. As this crop requires insect pollinators in order to set fruit, a thorough study of the effectiveness of its pollinators and a description of floral rewards would benefit growers by helping to refine pollination management strategies.

1.3 Assessing pollinator performance

There are multiple criteria for determining the potential of a species as a pollinator of a particular crop (Bosch and Kemp, 2002). Firstly, a candidate species needs to be active at the same time and under the same conditions as the target crop is blooming. For Haskap, this means the adult pollinator needs to be active very early in the spring, and also needs to be tolerant of the cold temperatures that are likely at this time

4 of year. Secondly, there needs to be potential for a sufficiently high density of foragers to pollinate the crop. This is one of the main benefits of using A. mellifera colonies for commercial pollination, as a single hive can contain thousands of workers. This contrasts with many native solitary species, which exist in naturally low densities, although some may be managed to artificially raise densities to very high levels (e.g. Megachile rotundata, Osmia lignaria). Next, the candidate species needs to be an effective

(successfully transfers pollen from anther to stigma) and efficient (contributes to high levels of fruit set) pollinator of the target crop. This can be deduced using proxies such as visit duration or stigma contact, but is best determined by direct measures of single visit pollen deposition along with behavioral variables such as visitation patterns and rate

(King et al., 2013; Ne’eman et al., 2010). Lastly, the species has to have a preference for the target crop, choosing to visit it even in the presence of alternative forage.

Currently, A. mellifera is the most common species managed to supplement

Haskap pollination. Most orchardists that grow Haskap in Saskatchewan either maintain their own apiary, or partner with local apiarists to bring in A. mellifera colonies during flowering. Apis mellifera colonies are often a practical solution to pollination deficits, as they are commonly available and easily transportable, making it possible to apply high densities of pollinators where needed (Klein et al., 2007). They also appear to be attracted to Haskap flowers, which can be an abundant early season resource for A. mellifera as well as for native nectar and pollen-seeking insects. However, they are not very cold tolerant, which may be problematic for Haskap as temperatures during flowering are often well below 20˚C, the optimal foraging temperature for A. mellifera

(Tan et al., 2012).

Osmia lignaria Say, the blue orchard bee, is another managed species that is being considered by some growers in Saskatchewan as a potential pollinator for Haskap.

Osmia lignaria is a solitary cavity nesting species that will also nest in artificial nest boxes, and they can be introduced into orchards in relatively high numbers (Bosch and

Kemp, 2001). They show a particular preference for rosaceous tree fruits, and are effective in supplementing pollination of several crops, including apple and cherry

(Bosch and Kemp, 1999; Torchio, 1976, 1985). As a native species to Canada, they are active in cooler temperatures than A. mellifera, and their active period naturally coincides with the flowering of Haskap. Osmia lignaria is active for about 6 weeks in the spring, which also means they do not require year-round forage, unlike A. mellifera and

Bombus spp. colonies which require nectar and pollen throughout the growing season.

Wild Bombus are one of the first groups of bees to emerge in the spring and may be the natural pollinators of many Lonicera species in the wild, including Haskap, and therefore may be one of the best pollinators of the crop (Hummer et al., 2012). They are very cold tolerant and are able to fly in temperatures as low as 0˚C (Heinrich, 2004); as

Haskap flowers will survive even through sub-zero temperatures, Bombus may be integral to pollinating Haskap during cold and unpredictable springs. However, individual Bombus active in early spring are solitary queens, whose worker brood have not yet matured. This limits the potential density of foragers available for pollination early in the season. Additionally, the density of queens in the spring at a given site is dependent on the size of colonies produced the previous fall (i.e., when the new queens are finally produced), which in turn means that forage needs to be available for them year-round.

Whereas all three of these groups of bees have potential to be important contributors to Haskap pollination, none of them have been formally evaluated for this crop. Importantly, the floral preferences of these groups need to be analyzed, and the effectiveness and efficiency of each taxon determined.

1.4 Study objectives

The overall purpose of my research was to build knowledge about the interaction between cultivated Haskap and its insect pollinators. A detailed knowledge of the pollination of Haskap will contribute to the field of pollination biology as a whole, as well as providing important information for Haskap producers in order to maintain high yields.

The specific objectives were to:

1. Determine the performance of A. mellifera, O. lignaria, and wild Bombus as

pollinators of Haskap, using single visit pollen deposition, visit symmetry

(i.e. the likelihood of visiting both flowers in the inflorescence), visit

duration, activity levels, and pollen preferences. Determining the importance

of visit symmetry also necessitated the determination of the effect of different

pollination treatments on fruit development, and minimum pollination

requirements (using seed number and pollen viability) were determined to

relate pollen deposition to fruit set. This research has been published as Frier

et al. (2016).

2. Determine floral longevity, nectar dynamics, pollen release, and stigma

receptivity of Haskap flowers in order to learn more about the plant’s

reproductive strategy and its associated animal pollinators. I asked four major

questions:

a. How long are the flowers open for, and is floral longevity affected by

pollination?

b. When is nectar produced, and is total nectar production affected by

nectar removal?

c. When is pollen released from the anthers?

d. How early in anthesis does the stigma become receptive?

This research has been written as a manuscript and has been submitted to

Journal of Pollination Biology.

References

Bors, B., 2008. Growing Haskap in Canada [WWW Document]. U S Fruit Progr. URL http://www.fruit.usask.ca/articles/growinghaskapinCanada.pdf (accessed 9.21.13). Bors, B., Thomson, J., Sawchuk, E., Reimer, P., Sawatzky, R., Sander, T., Kaban, T., Gerbrandt, E., Dawson, J., 2012. Haskap berry breeding and production - final report. Bosch, J., Kemp, W., 1999. Exceptional cherry production in an orchard pollinated with blue orchard bees. Bee World 80, 163–173. Bosch, J., Kemp, W., 2001. How to Manage the Blue Orchard Bee As an Orchard Pollinator, Systematics. Sustainable Agriculture Network. Bosch, J., Kemp, W.P., 2002. Developing and establishing bee species as crop pollinators: the example of Osmia spp. (Hymenoptera: Megachilidae) and fruit trees. Bull. Entomol. Res. 92, 3–16. doi:10.1079/BER2001139 Bożek, M., 2007. Pollen productivity and morphology of pollen grains in two cultivars of honeyberry (Lonicera kamtschatica (Sevast.) Pojark.). Acta Agrobot. 60, 73–77. Bożek, M., 2012. The effect of pollinating insects on fruiting of two cultivars of Lonicera caerulea L. J. Apic. Sci. 56, 5–11. doi:10.2478/v10289-012-0018-6

Bożek, M., Wieniarska, J., 2006. Biologia kwitnienia i wydajność cukrowa kwiatów dwóch odmian Lonicera kamtschatica (Sevast.) Pojark / Blooming biology and sugar efficiency of two cultivars of Lonicera kamtschatica (Sevast.) Pojark. Acta Agrobot. 59, 177–182. Chaovanalikit, A., Thompson, M.M., Wrolstad, R.E., 2004. Characterization and quantification of anthocyanins and polyphenolics in blue honeysuckle (Lonicera caerulea L.). J. Agric. Food Chem. 52, 848–852. doi:10.1021/jf030509o Frier, S.D., Somers, C.M., Sheffield, C.S., 2016. Comparing the performance of native and managed pollinators of Haskap (Lonicera caerulea: Caprifoliaceae), an emerging fruit crop. Agric. Ecosyst. Environ. 219, 42-48. doi:10.1016/j.agee.2015.12.011 Heinrich, B., 2004. Bumblebee Economics, 2 edition. ed. Harvard University Press. Hummer, K.E., Pomper, K.W., Postman, J., Graham, C.J., Stover, E., Mercure, E.W., Aradhya, M., Crisosto, C.H., Ferguson, L., Thompson, M.M., Byers, P., Zee, F., 2012. Fruit Breeding. doi:10.1007/978-1-4419-0763-9 King, C., Ballantyne, G., Willmer, P.G., 2013. Why flower visitation is a poor proxy for pollination: measuring single-visit pollen deposition, with implications for pollination networks and conservation. Methods Ecol. Evol. 4, 811–818. Klein, A.-M., Vaissière, B.E., Cane, J.H., Steffan-Dewenter, I., Cunningham, S. a, Kremen, C., Tscharntke, T., 2007. Importance of pollinators in changing landscapes for world crops. Proc. R. Soc. B 274, 303–13. Miller, C., Dulac, E., Miller, X., 2011. 5. LONICERA Linnaeus, Sp. Pl. 1: 173. 1753., in: Flora of China. pp. 620–641. Ne’eman, G., Jürgens, A., Newstrom-Lloyd, L., Potts, S.G., Dafni, A., 2010. A framework for comparing pollinator performance: effectiveness and efficiency. Biol. Rev. Camb. Philos. Soc. 85, 435–51. doi:10.1111/j.1469-185X.2009.00108.x Plekhanova, M.N., 2000. Blue honeysuckle (Lonicera caerulea L.) - A new commercial berry crop for temperate climate: genetic resources and breeding. Acta Hortic. 538, 159–164. Raudsepp, P., Anton, D., Roasto, M., Meremäe, K., Pedastsaar, P., Mäesaar, M., Raal, A., Laikoja, K., Püssa, T., 2013. The antioxidative and antimicrobial properties of the blue honeysuckle (Lonicera caerulea L.), Siberian rhubarb (Rheum rhaponticum L.) and some other plants, compared to ascorbic acid and sodium nitrite. Food Control 31, 129–135. doi:10.1016/j.foodcont.2012.10.007

Rehder, A., 1903. Synopsis of the genus Lonicera. Missouri Bot. Gard. Annu. Rep. 1903, 27–232. Rehder, A., 1909. Note on the morphology of the fruit of Lonicera caerulea. Rhodora 11, 209–211. Rop, O., Řezníček, V., Mlček, J., Juríková, T., Balík, J., Sochor, J., Kramářová, D., 2011. Antioxidant and radical oxygen species scavenging activities of 12 cultivars of blue honeysuckle fruit. Hortic. Sci. 38, 63–70. Rupasinghe, H.P.V., Yu, L.J., Bhullar, K.S., Bors, B., 2012. Haskap (Lonicera caerulea): A new berry crop with high antioxidant capacity. Can. J. Plant Sci. 3, 1311–1317. doi:10.4141/CJPS2012-073 Svarcova, I., Heinrich, J., Valentova, K., 2007. Berry fruits as a source of biologically active compounds: the case of Lonicera caerulea. Biomed. Pap. 151, 163–174. Tan, K., Yang, S., Wang, Z.-W., Radloff, S.E., Oldroyd, B.P., 2012. Differences in foraging and broodnest temperature in the honey bees Apis cerana and A. mellifera. Apidologie 43, 618–623. doi:10.1007/s13592-012-0136-y Thompson, M.M., 2006. Introducing Haskap , Japanese blue honeysuckle. J. Am. Pomol. Soc. 60, 164–168. Torchio, P.F.., 1976. Use of Osmia lignaria Say (Hymenoptera : Apoidea , Megachilidae) as a pollinator in an apple and prune orchard. J. Kansas Entomol. Soc. 49, 475–482. Torchio, P.F., 1985. Field experiments with the pollinator species, Osmia lignaria propinqua Cresson, in apple orchards: V (1979-1980), methods of introducing bees, nesting success, seed counts, fruit yields (Hymenoptera: Megachilidae). J. Kansas Entomol. Soc. 58, 448–464.

CHAPTER 2: Comparing the performance of native and managed pollinators of

Haskap (Lonicera caerulea: Caprifoliaceae), an emerging fruit crop

1. Introduction

Pollination services for insect pollinated crops can be provided by the introduction of managed bee species, or by wild populations of native pollinators. Apis mellifera L. (Hymenoptera: Apidae), the honey bee, is the most widely used managed pollinator, and many crops directly depend on its use (Klein et al., 2007; McGregor,

1976). As social, generalist, and commercially available foragers, A. mellifera can be a practical solution for avoiding pollination deficits when native pollinators are scarce

(Klein et al., 2007). However, it is well known that A. mellifera workers are inefficient pollinators of some crops, and in many cases alternative managed or wild species may do a better job [e.g. reviewed by Klein et al., 2007; Westerkamp and Gottsberger, 2000; almond (Bosch and Blas, 1994); blueberry (Javorek et al., 2002); coffee (Klein et al.,

2003); raspberry and blackberry (Cane, 2005); tomatoes (Putra and Kinasih, 2014); cherry (Bosch and Kemp, 1999); pear (Monzón et al., 2004)]. Even when A. mellifera are effective pollinators, native bees increase crop yields and buffer against commercial colony losses (Garibaldi et al., 2011, 2013; Klein et al., 2003; Rader et al., 2013; Winfree et al., 2007). For many crops the performance of different species of pollinators is still unknown, and A. mellifera is often used by default even when they may not be the most effective or efficient option.

Haskap (Lonicera caerulea: Caprifoliaceae), also known as blue honeysuckle or honeyberry, is a temperate fruiting shrub that is visited by A. mellifera, Bombus spp. (i.e. bumble bees; Hymenoptera Apidae), and a variety of solitary bees (Bożek, 2012).

Haskap is native to northern regions of North America, Europe, and Asia, and it has recently gained popularity as a commercial crop in North America (Hummer et al.,

2012). Haskap production is increasing in North America due to potential health benefits of eating the fruit (e.g. Lefèvre et al., 2011; Rupasinghe et al., 2012; Svarcova et al.,

2007), and the slightly tart flavour of its blue berries, which are eaten fresh or used in a wide variety of food products (Hummer et al., 2012). However, very little is known about Haskap pollination biology, though it is known to be self-incompatible (Bors et al.,

2012; Hummer et al., 2012), and to require animal pollinators (Bożek, 2012). In Canada,

A. mellifera is the primary managed pollinator used for Haskap production, with most orchards receiving some level of pollination services from local apiaries. Recently,

Osmia lignaria Say (Hymenoptera: Megachilidae), known commonly as the blue orchard bee, has also been considered as a potential pollinator of Haskap. Native Bombus spp., solitary bees (e.g. Lasioglossum spp., Osmia spp. Halictus spp.), and syrphid flies

(Diptera: Syrphidae) are also regular visitors of Haskap in Canada; however, the pollination efficacy of both managed and wild floral visitors of Haskap is entirely unknown.

Haskap flowers have a pale yellow tubular corolla, less than 2.5cm in length, and grow in two flowered cymes that are typical of other Lonicera species. Many inflorescences are produced per branch and the flowers nod downwards. The flowering period for Haskap lasts 2 to 3 weeks and occurs very early in the spring before the plant leafs out, and the flowers are able to tolerate frosts down to at least -7˚C (Bors, 2008;

Hummer et al., 2012). Only a few insect pollinators can fly during periods of low temperature characteristic of spring at high latitudes, potentially suggesting a dependence on cold-tolerant pollinators such as Bombus spp., whose large queens can fly in temperatures as low as 0˚C (Heinrich, 2004). Fruit development in Haskap is unusual

13 compared to other crops; the bracteoles subtending the flowers have fused together around the two ovaries (Fig. 1a), and the Haskap berry (actually a multiple fruit) forms from these combined structures (Figs. 1b and 1c). The development of the Haskap fruit may depend on the successful pollination of both flowers in the inflorescence, so pollinator behaviour (e.g. how likely a species is to visit both flowers) may be especially important for the production of high quality fruit (Woodcock et al., 2013).

The purpose of this study was to assess the pollination performance of commercially managed A. mellifera and O. lignaria relative to wild Bombus spp. on

Haskap crops in Saskatchewan, Canada. Specifically, we compared single visit pollen deposition, visit symmetry, visit duration, and pollinator density. In addition, we analyzed the pollen load composition from female O. lignaria; although A. mellifera and

Bombus spp. are known to forage on Haskap (Bożek, 2012), it is unclear whether O. lignaria show a preference for the crop. We hypothesized that because Bombus spp. are active in lower temperature ranges that correspond with flowering, they are likely the most effective and efficient pollinators of Haskap. We predicted that A. mellifera, which is a non-native species that prefers warm weather that is unlikely to co-exist with Haskap under natural conditions, would be less effective and efficient. Osmia lignaria is known to be an effective pollinator of other fruit crops (Bosch and Kemp, 2001), and is a native

North American species that is well adapted to cool temperatures, so we expected them to be similar to Bombus spp. in performance. These taxa have never been assessed as pollinators of Haskap, and the information obtained from this study will help Haskap orchard managers to implement pollination strategies that optimize fruit yield in a sustainable manner.

2. Methods

2.1 Study site

Our research was conducted in the spring of 2014 within a 30 hectare organic

Haskap orchard in Birch Hills, Saskatchewan, Canada (52.97 N, -105.42 W). The study orchard was bordered on two sides by a narrow shelter-belt of trees. The surrounding area was largely agricultural and consisted mostly of crops such as canola, alfalfa, and wheat. Plants in the area that flowered concurrently to Haskap were primarily willow

(Salix, Salicaceae) and dandelion (Taraxacum, Asteraceae). Various species of insects were frequently observed visiting Haskap flowers within the orchard, including over 10 species of Bombus (Appendix A, supplementary material). The orchard was established in 2008 and grows four Haskap cultivars developed in the University of Saskatchewan’s fruit breeding program. Our research focused on the varieties ‘Tundra’ and ‘Indigo

Gem’, which were chosen because they were well represented within the orchard by mature plants. Each row within our study area contained approximately 400 plants of the same cultivar, spaced 1 metre apart, with adjacent rows being different cultivars and a row of pollinizers (a distantly related cultivar that provides a compatible source of pollen to the desired cultivars) located every 5th row. The cultivars ‘Tundra’ and ‘Indigo Gem’ were represented by a total of 13 and 14 rows, respectively.

For the purpose of our experiments, 12 full-strength colonies of A. mellifera provided by a local apiarist were placed on the eastern edge of the orchard. In addition,

12 nesting boxes of O. lignaria were obtained from Mason Bee Central (Black Creek,

B.C, Canada) for our experiments, with 220 bees (2:1 female to male ratio) in each.

2.2 Single visit deposition

To assess pollen deposition by each pollinator taxa, we covered branches of unopened flowers with pollinator exclusion bags. Once the flowers had opened we removed the whole branch from the bush and placed it into a single-stem flower tube containing water and 20-20-20 NPK fertilizer. This branch was then presented to foraging bees by placing the unvisited flowers just next to a flower that a bee was visiting. Immediately following a visit, the inflorescence was removed from the branch and the stigmas of both flowers mounted in gelatin-fuchsine (following Dafni et al.

2005). Unvisited flowers from the same branch were removed at random intervals and used as controls to account for autogamous pollen that contacted the stigma by means other than a bee visit (e.g., wind, handling). The amount of pollen deposited by a bee on a single visit (single visit deposition, hereafter SVD) is the difference between the pollen from unvisited controls (internal self-pollination) and the total amount of pollen found on visited stigmas.

The slide-mounted stigmas were digitally photographed using a Canon EOS 5D camera with a Canon macro photo lens MP-E 65mm 1:2.8 1-5x and a Tamron SP AF tele-converter 140F –CA 1.4x with dynalite MH2050 lights on a studio-pro dynalite

1600 2/s flash generator. Multiple images at different focal depths (each 0.12mm) were taken and then stacked (Helicon Focus 5.3.3) to create one image with all pollen grains in focus. The photographs were then imported to ImageJ (Rasband, 2014) version 1.48 and individual Haskap pollen grains were counted manually using the “cell counter” plug-in.

2.3 Visit symmetry and duration

Visit symmetry was derived alongside data gathered for SVD. For every successful visit, we recorded whether the bee visited 1 or 2 flowers in the inflorescence.

Visit duration was measured by observing foragers as they visited Haskap inflorescences. The length of the observation (in seconds) was measured, and the number of inflorescences visited during the observation was tallied using a hand-held counter.

Timing began when the bee was first observed landing on a single inflorescence, and ended when the bee moved out of sight of the observer. The number of inflorescences visited per time interval was then used to calculate the average visit duration for each observation, and this final measure represents both search and forage time per inflorescence.

2.4 Pollinator density

Transects were set up at three locations within the study site. Each transect was placed between two adjacent rows of Haskap, with cultivar ‘Tundra’ on one side and

‘Indigo Gem’ on the other. The transects were surveyed daily every two hours, starting from the beginning of flowering (24-May-2015) and finishing once the target cultivars were near the end of their flowering period (10-Jun-2015). Surveys were not performed when it was raining. Air temperature and average wind speed were measured at the beginning of each survey using a Kestrel 2000 pocket weather meter. The surveys were conducted by scan sampling (following Vaissière et al. 2011); a total of 400 inflorescences were sampled per cultivar.

2.5 Pollen load composition of Osmia lignaria

To obtain more information about the pollen preferences of O. lignaria, we captured females (i.e. the pollen foragers) as they returned to their nests. We used a cotton swab to remove pollen from their abdomen before releasing them; the pollen was then stored dry in a micro-centrifuge vial (after Sheffield et al. 2008). Pollen-loads were sampled from 12 different nests within the orchard on 4 different sampling dates throughout the flowering period.

The cotton swabs with pollen samples were put into 70% ethanol and shaken or mixed using a spatula to disperse the pollen. The ethanol-pollen mixture was filtered through Whatman #1 filter paper in a Buchner funnel, dried, and slide-mounted using fuchsine gelatine following Dafni et al. (2005). Slides were photographed using the same method used for SVD. Photographs were then overlaid with a 5x5 grid using Adobe

Photoshop CS5 Extended (Version 1.2.4 x64), and all the grains within at least 5 grid spaces were counted. At least 300 pollen grains were counted per sample. Pollen grains were identified to genus using published references based on size and morphology

(Bassett, 1978; Bożek, 2007; Crompton, 1993).

2.6 Statistical analysis

The data from all experiments were analyzed using a generalized linear model

(GLM) approach, with appropriate error structures and testing for over-dispersion. Pollen counts from visited flowers were fitted to a negative binomial distribution, with number of pollen grains as the response and pollinator taxon as the predictor, and using the unvisited controls as the reference group. SVD was then inferred using the estimates ±

SE from the resulting model and compared to the minimum pollen requirement

(Appendix B, supplementary material). Visit symmetry and visit duration were fitted to a gamma and binomial distribution, respectively. For visit symmetry, the response variable was a binary response where 1 or 2 flowers were visited, with pollinator taxon as the predictor. For visit duration, the response variable was average visit time and pollinator taxon was the predictor. Data from the pollinator density surveys were checked for autocorrelation, and none was observed, although the data were over-dispersed, and so were fitted to a quasi-poisson distribution using pollinator density as the response variable, and temperature, time of day, cloud cover, wind speed and density of other taxa as predictor variables. Due to low sample sizes, O. lignaria were excluded from the density analysis. The proportions of each pollen type found in the pollen loads of O. lignaria females were also over-dispersed and were fitted to a quasi-binomial distribution and analyzed separately with the proportion of each pollen type as the response variable and date as the predictor variable. Chi-squared test statistics for each overall model were obtained by an analysis of variance against the null model, and multiple comparisons between groups were performed for significance using Tukey’s method. All statistical analyses were performed in R using the basic stats (R Core Team

2014), MASS (Venables and Ripley, 2002), and multcomp (Hothorn et al., 2008) packages.

3. Results

3.1 Single visit deposition

Much variation was observed in the amount of Haskap pollen found on the stigmas, regardless of treatment group [A. mellifera (mean=85 sd=64, df=141, range=0-

390), Bombus spp. [mean 99, sd=73, df=122, range=0-356), O. lignaria (mean=107, sd=92, df=16, range=22-327), control (mean=72.1, sd=64, df=96, range=0-296)]. The mean amount of pollen deposited on stigmas visited by each taxon was higher than the unvisited control group, but only significantly so for stigmas visited by Bombus spp.

(Table 1). The estimated mean SVD levels for Bombus spp. and O. lignaria surpassed the minimum pollination requirements of 24 pollen grains per flower; whereas those for

A. mellifera did not (Fig. 2).

3.2 Visit symmetry and duration

Apis mellifera were slightly less likely than Bombus spp. or O. lignaria to visit both flowers; 31/88 (35%) of A. mellifera visits included both flowers, versus 44/91

(48%) and 6/11 (55%) for Bombus spp. and O. lignaria, respectively. However, these differences were not significant (GLM; Χ2=3.85, df=2, p=0.146). Visit duration differed between taxa (GLM; Χ2=42.90, df=2, p=<0.001) and was lowest in Bombus spp.

(mean=6s, SD=3s, n=107), highest in A. mellifera (mean=15s, SD=7s, n=82), with O. lignaria intermediate between these two (mean=10s, SD=4s, n=24).

3.3 Pollinator density

In total, 99 surveys were performed over a period of 18 days. The mean temperature recorded from all surveys was 18.0˚C (range=8.3 to 30.5˚C), with mean wind speeds of 8.9km/h (range=0.0 to 30.8km/h). Apis mellifera and Bombus spp. were regularly observed foraging within the orchard, however, foraging O. lignaria were rarely observed, likely due to low retention of nesting females within the orchard (at maximum, 250 nesting females were observed in the nests). This made the likelihood of detection within the orchard very low, and therefore O. lignaria was excluded from this analysis. During peak activity, A. mellifera workers were much more abundant than

Bombus spp. (Fig. 3). However, A. mellifera abundance fluctuated much more throughout the day, while Bombus spp. numbers stayed relatively constant. Apis mellifera workers were rarely observed foraging below about 13˚C, whereas Bombus spp. foragers were present below 10˚C (Fig. 4). A negative binomial regression model identified time of day as being the only significant predictor variable for density of foraging Bombus spp., whereas temperature was the only significant variable for density of A. mellifera (Table 2).

3.4 Pollen load composition of Osmia lignaria

Four different pollen types were observed in pollen loads sampled from female

O. lignaria. Three were identified as belonging to Lonicera caerulea, Salix sp., and

Taraxacum, respectively; one was unable to be identified. Overall, about 63% of the pollen samples collected from O. lignaria contained no Haskap pollen, and 27% of the samples contained only trace amounts (>5%) of Haskap. The most common pollen found

21 in the samples was from the genus Salix, with 85% of the samples analyzed containing more than 90% Salix pollen. We found no temporal differences in the source of pollen.

4. Discussion

Although SVD is considered one of the most direct and practical methods of assessing pollinator effectiveness (King et al., 2013), the results of SVD analysis in

Haskap are difficult to interpret. The orientation of the anthers in relation to the stigma in individual flowers means that large amounts of self-pollen can be transferred to the stigma, making it difficult to determine if pollen was deposited by the pollinators or via internal (i.e. intra-floral) self-pollination. Removal of the anthers prior to dehiscence is a possible solution, though manipulation of unopened flowers may affect the attractiveness of the flowers and therefore the visitation and behaviour of pollinators, which could bias the results (e.g. Rademaker et al., 1997). Although previous studies have assumed that species without significantly greater SVD than unvisited controls are ineffective pollinators (King et al., 2013), we hesitate to draw the same conclusion. Based on high fruit yields observed in this orchard, we can confidently say that effective cross pollination of Haskap occurred, and our results suggest that Bombus spp. are the best pollinators of those we studied, but we do not rule out effective pollination by A. mellifera or O. lignaria. Rather, it is likely that statistical analysis of SVD on Haskap has limited power due to a large variation in the amount of self-pollination that occurs, and future studies on Haskap and similar flowers should consider looking at fruit set rather than pollen deposition.

The high level of self-pollination we observed in Haskap has implications for the monitoring of Haskap pollination as well as direction for future research. Background pollination assessments, which examine the average amount of pollen on stigmas of a flowering crop, are often done to determine whether poor fruit yield is a result of poor pollination or from compatibility issues. If plenty of pollen is found on the stigmas of flowering plants but yields remain low, it might be assumed that the problem is with incompatibility between cultivars and/or insufficient pollinizer abundance. However,

Haskap flowers may have a lot of pollen on the stigma even in the absence of sufficient pollinator services, due to intra-floral self-pollination. If levels of self-pollination are not accounted for, this could lead to the erroneous assumption that there are problems with compatibility. For example, background pollination levels were assessed at our study site in 2013 (Frier, unpublished data), and 83% of the stigmas analyzed were found to have the minimum pollen requirement, given maximum seed number and average pollen viability (Appendix B, supplementary material). However, when we factor in the amount of intra-floral self-pollination that occurs, only 53% of the flowers were sufficiently pollinated. In addition, as Haskap is self-incompatible (Bors et al., 2012; Hummer et al.,

2012), these high levels of self-pollination could potentially result in pollen clogging of the stigma, a type of sexual interference that can result in reduced reproductive success

(Barrett, 2002; Bertin and Sullivan, 1988; Broyles and Wyatt, 1993; Galen et al., 1989;

Waser and Price, 1991).

Despite the difficulties in analyzing SVD, as discussed above, Bombus spp. appear to be the most effective and efficient pollinators of Haskap on an individual basis.

Bombus spp. have significantly higher SVD levels than the other taxa assessed here, and

23 likely require no more than two visits on average in order to fully pollinate a Haskap flower, compared to A. mellifera which almost certainly require multiple visits (Fig. 2).

Individual Bombus spp. are also able to visit more inflorescences per day than individual

A. mellifera workers due to their low visit duration times, and Bombus spp. density (and therefore their assumed individual activity levels) is independent of temperature within the range experienced during Haskap flowering. This increased effectiveness and efficiency of wild Bombus pollinators may compensate for their lower density compared to honey bees, and their cold tolerance may make them essential for ensuring pollination during cool spring temperatures (Fig. 4). This supports our hypothesis that Haskap, particularly the early blooming varieties, may rely substantially on pollination by

Bombus spp. in the wild and may have developed specialized characteristics to capitalize on this relationship. Indeed, Bombus spp. have been recorded as visitors of Haskap elsewhere in its range (Bożek, 2012; Hummer et al., 2012; Matsumura et al., 2004), although more thorough study of its floral visitors throughout the distribution of Haskap is necessary.

Apis mellifera were the least effective pollinators of Haskap when considered individually. They have lower SVD levels than Bombus spp. or O. lignaria, spend three times as long as Bombus spp. per floral visit, are less active in low temperatures, and may be less likely than both Bombus spp. and O. lignaria to visit both flowers in the inflorescence. However, during fair weather, A. mellifera workers were observed to be the most abundant visitor to Haskap. As a result, the large pollinator force available in commercial hives may mean that overall they contribute more than any other species to

Haskap pollination, particularly in orchards with surrounding habitat which is poor for

24 native bees (Kremen et al., 2004; Ricketts et al., 2008). This concept of “brute force” pollination (Morse, 1991) suggests that even though individuals of A. mellifera may not be very effective pollinators of some crops, their high density and the mobility of colonies makes them an ideal solution for pollination deficits. However, we must emphasize again the low activity levels of A. mellifera in cooler weather; although the introduction of commercial hives may have a positive impact on yield, the total effect will be dependent on highly variable weather conditions.

In contrast, O. lignaria performed well in several of the analyses conducted here, but they are poor overall pollinators of Haskap in field conditions. Visit rate and symmetry exceeded that of A. mellifera, and although our ability to draw conclusions about SVD in O. lignaria was limited by low sample size, it appears that they may deposit as much or more pollen than Bombus spp. (Fig. 2). We speculate that this might be due to the scopa (the pollen carrying structure), being located on the underside of the abdomen, a characteristics shared by most Megachilidae (Michener, 2000) that may enhance the probability of the pollen load contacting the stigma. This could suggest also that wild Osmia, which were not included in this analysis, could be very effective pollinators of Haskap. However, analyses of pollen loads from foraging O. lignaria reveal that this species has a very low affinity for Haskap, and even when they are nesting in the middle of an orchard in full bloom they prefer alternative pollen sources, especially Salix. Osmia lignaria will occasionally visit Haskap for nectar and rarely for pollen collection, as evidenced by Haskap grains found in some pollen load samples as well as direct observations of foragers on Haskap flowers, but their contribution to

Haskap pollination is likely to be negligible overall in field conditions. Efforts are

25 currently underway at the University of Saskatchewan, Canada, to develop Haskap as a greenhouse crop; it is possible that O. lignaria may be effective pollinators under these circumstances, where no other forage is available, but further study into their SVD should be done to confirm their status as a pollinator of Haskap.

5. Conclusions

Our results suggest that Bombus queens are highly effective as Haskap pollinators when compared to currently managed A. mellifera workers and female O. lignaria. While we are unable to make a direct link from pollination performance to fruit set, it is likely that Haskap production would benefit from high populations of wild

Bombus spp. Importantly, because it is the solitary Bombus queens, and not workers, that are visiting Haskap, sufficient numbers of Bombus available for pollination in the spring is dependent on the production of large, healthy colonies with new queen production late in the summer and fall. Haskap production should be coupled with later-flowering crops, and uncultivated areas around orchards, such as shelter belts and drainage ditches, and areas between rows could be planted with bee-friendly ground cover plants such as clover, which will provide year-round forage and enhance crop visitation by wild bees

(Garibaldi et al., 2011; Ricketts et al., 2008). In Haskap orchards where the surrounding habitat is poor for bees or where the orchard is particularly large, resulting in low yield,

A. mellifera colonies could be introduced to supplement pollination by native bees.

Mutually beneficial relationships could arise between Haskap growers and apiarists, as

Haskap provides rich and abundant forage for bees early in the year when little else is available – even when hives are not placed in the orchard, individual A. mellifera

26 foragers have been observed to travel several kilometres to forage on it. It is important to note, however, that A. mellifera will only be effective pollinators of Haskap during good weather, and promoting native Bombus may be critical for ensuring good pollination during inclement conditions. Conversely, O. lignaria appear to be inefficient pollinators of Haskap due to their preference for alternative pollen sources such as willow, and we do not recommend that they be used for pollination of Haskap.

References

Barrett, S.C.H., 2002. Sexual interference of the floral kind. Heredity (Edinb). 88, 154– 159. Bassett, J.I., 1978. An atlas of airborne pollen grains and common fungus spores of Canada. Research Branch, Canadian Department of Agriculture, Ottawa, Ontario. Bertin, R.I., Sullivan, M., 1988. Pollen interference and cryptic self-fertility in Campsis radicans. Am. J. Bot. 75, 1140–1147. Bors, B., 2008. Growing Haskap in Canada [WWW Document]. U S Fruit Progr. URL http://www.fruit.usask.ca/articles/growinghaskapinCanada.pdf (accessed 9.21.13). Bors, B., Thomson, J., Sawchuk, E., Reimer, P., Sawatzky, R., Sander, T., Kaban, T., Gerbrandt, E., Dawson, J., 2012. Haskap berry breeding and production - final report. Bosch, J., Blas, M., 1994. Foraging behaviour and pollinating efﬁciency of Osmia cornuta and Apis mellifera on almond (Hymenoptera: Megachilidae, Apidae). Appl. Entomol. Zool. 29, 1–9. Bosch, J., Kemp, W., 2001. How to Manage the Blue Orchard Bee As an Orchard Pollinator, Systematics. Sustainable Agriculture Network. Bosch, J., Kemp, W., 1999. Exceptional cherry production in an orchard pollinated with blue orchard bees. Bee World 80, 163–173. Bożek, M., 2012. The effect of pollinating insects on fruiting of two cultivars of Lonicera caerulea L. J. Apic. Sci. 56, 5–11. doi:10.2478/v10289-012-0018-6 Bożek, M., 2007. Pollen productivity and morphology of pollen grains in two cultivars of honeyberry (Lonicera kamtschatica (Sevast.) Pojark.). Acta Agrobot. 60, 73–77.

Broyles, S., Wyatt, R., 1993. The consequences of self-pollination in Asclepias exaltata, a self-incompatible milkweed. Am. J. Bot. 80, 41–44. Cane, J., 2005. Pollination potential of the bee Osmia aglaia for cultivated red raspberries and blackberries (Rubus: Rosaceae). Hortscience 40, 1705–1708. Crompton, C.W., 1993. Pollen grains of Canadian honey plants. Agriculture Canada, Ottawa, Ontario. Dafni, A., Kevan, P.G., Husband, B.C., 2005. Practical Pollination Biology. Enviroquest Ltd., Cambridge, Ontario. Galen, C., Gregory, T., Galloway, L., 1989. Costs of self-pollination in a self- incompatible plant, Polemonium viscosum. Am. J. Bot. 76, 1675–1680. Garibaldi, L.A., Steffan-Dewenter, I., Kremen, C., Morales, J.M., Bommarco, R., Cunningham, S. A, Carvalheiro, L.G., Chacoff, N.P., Dudenhöffer, J.H., Greenleaf, S.S., Holzschuh, A., Isaacs, R., Krewenka, K., Mandelik, Y., Mayfield, M.M., Morandin, L. A, Potts, S.G., Ricketts, T.H., Szentgyörgyi, H., Viana, B.F., Westphal, C., Winfree, R., Klein, A.M., 2011. Stability of pollination services decreases with isolation from natural areas despite honey bee visits. Ecol. Lett. 14, 1062–72. Garibaldi, L.A., Steffan-Dewenter, I., Winfree, R., 2013. Wild pollinators enhance fruit set of crops regardless of honey bee abundance. Science 339, 1608–1611. Heinrich, B., 2004. Bumblebee Economics, 2 edition. ed. Harvard University Press. Hothorn, T., Bretz, F., Westfall, P., 2008. Simultaneous inference in general parametric models. Biometrical J. 50, 346–363. Hummer, K.E., Pomper, K.W., Postman, J., Graham, C.J., Stover, E., Mercure, E.W., Aradhya, M., Crisosto, C.H., Ferguson, L., Thompson, M.M., Byers, P., Zee, F., 2012. Fruit Breeding. doi:10.1007/978-1-4419-0763-9 Javorek, S.K., Mackenzie, K.E., Vander Kloet, S.P., 2002. Comparative pollination effectiveness among bees (Hymenoptera: Apoidea) on lowbush blueberry (Ericaceae: Vaccinium angustifolium). Ann. Entomol. Soc. Am. 95, 345–351. King, C., Ballantyne, G., Willmer, P.G., 2013. Why flower visitation is a poor proxy for pollination: measuring single-visit pollen deposition, with implications for pollination networks and conservation. Methods Ecol. Evol. 4, 811–818. Klein, A.-M., Steffan-Dewenter, I., Tscharntke, T., 2003. Fruit set of highland coffee increases with the diversity of pollinating bees. Proc. R. Soc. BB 270, 955–61.

Klein, A.-M., Vaissière, B.E., Cane, J.H., Steffan-Dewenter, I., Cunningham, S. a, Kremen, C., Tscharntke, T., 2007. Importance of pollinators in changing landscapes for world crops. Proc. R. Soc. B 274, 303–13. Kremen, C., Williams, N.M., Bugg, R.L., Fay, J.P., Thorp, R.W., 2004. The area requirements of an ecosystem service: crop pollination by native bee communities in California. Ecol. Lett. 7, 1109–1119. Lefèvre, I., Ziebel, J., Guignard, C., Sorokin, A., Tikhonova, O., Dolganova, N., Hoffmann, L., Eyzaguirre, P., Hausmann, J.-F., 2011. Evaluation and comparison of nutritional quality and bioactive compounds of berry fruits from Lonicera caerulea, Ribes L. species and Rubus idaeus grown in Russia. J. Berry Res. 1, 159–167. doi:10.3233/BR-2011-017 Matsumura, C., Yokoyama, J., Washitani, I., 2004. Invasion status and potential ecological impacts of an invasive alien bumblebee, Bombus terrestris L. (Hymenoptera: Apidae) naturalized in Southern Hokkaido, Japan. Glob. Environ. Res. 8, 51–66. McGregor, S., 1976. Insect pollination of cultivated crop plants. U.S.D.A. Agriculture Handbook No. 496. Michener, C., 2000. The bees of the world. John Hopkins University Press, Baltimore, MD. Monzón, V., Bosch, J., Retana, J., 2004. Foraging behavior and pollinating effectiveness of Osmia cornuta (Hymenoptera: Megachilidae) and Apis mellifera (Hymenoptera: Apidae) on “Comice” pear. Apidologie 35, 575–585. Morse, R.A., 1991. Honeybees forever. Trends Ecol. Evol. (Personal Ed. 6, 337–338. doi:10.1016/0169-5347(91)90043-W Putra, R.E., Kinasih, I., 2014. Efficiency of local Indonesia honey bees (Apis cerana L.) and stingless bee (Trigona iridipennis) on tomato (Lycopersicon esculentum Mill) pollination. Pakistan J. Biol. Sci. 1, 86-91. R Core Team (2014) R: A language and environment for statistical computing. Rademaker, M., Jong, T.J. De, Klinkhamer, P.G.L., 1997. Pollen dynamics of bumblebee visitation on Echium vulgare. Funct. Ecol. 11, 554–563. Rader, R., Reilly, J., Bartomeus, I., Winfree, R., 2013. Native bees buffer the negative impact of climate warming on honey bee pollination of watermelon crops. Glob. Chang. Biol. 19, 3103–3110. Rasband, W.S., 2014. ImageJ version 1.49.

Ricketts, T.H., Regetz, J., Steffan-Dewenter, I., Cunningham, S. A., Kremen, C., Bogdanski, A., Gemmill-Herren, B., Greenleaf, S.S., Klein, A.M., Mayfield, M.M., Morandin, L. A., Ochieng, A., Viana, B.F., 2008. Landscape effects on crop pollination services: Are there general patterns? Ecol. Lett. 11, 499–515. Rupasinghe, H.P.V., Yu, L.J., Bhullar, K.S., Bors, B., 2012. Haskap (Lonicera caerulea): A new berry crop with high antioxidant capacity. Can. J. Plant Sci. 3, 1311–1317. doi:10.4141/CJPS2012-073 Sheffield, C.S., Westby, S.M., Smith, R.F., Kevan, P.G., 2008. Potential of bigleaf lupine for building and sustaining Osmia lignaria populations for pollination of apple. Can. Entomol. 140, 589–599. Svarcova, I., Heinrich, J., Valentova, K., 2007. Berry fruits as a source of biologically active compounds: the case of Lonicera caerulea. Biomed. Pap. 151, 163–174. Vaissière, B., Freitas, B., Gemmill-Herren, B., 2011. Protocol to detect and assess pollination deficits in crops: a handbook for its use. Food and Agriculture Organization of the United Nations, Rome.

Venables, W.N., Ripley, B.D., 2002. Modern Applied Statistics with S, 4th ed. Springer, New York.

Waser, N., Price, M., 1991. Reproductive costs of self-pollination in Ipomopsis aggregata (Polemoniaeae): are ovules usurped? Am. J. Bot. 78, 1036–1043. Westerkamp, C., Gottsberger, G., 2000. Diversity pays in crop pollination. Crop Sci. 40, 1209–1222. Winfree, R., Williams, N.M., Dushoff, J., Kremen, C., 2007. Native bees provide insurance against ongoing honey bee losses. Ecol. Lett. 10, 1105–1113. Woodcock, B. A., Edwards, M., Redhead, J., Meek, W.R., Nuttall, P., Falk, S., Nowakowski, M., Pywell, R.F., 2013. Crop flower visitation by honeybees, bumblebees and solitary bees: Behavioural differences and diversity responses to landscape. Agric. Ecosyst. Environ. 171, 1–8. doi:10.1016/j.agee.2013.03.005

Figure 1. Haskap (Lonicera caerulea L.) flowers and fruits. (a) A typical Haskap

inflorescence. The two perfect flowers each have five petals, five stamens and

one stigma; the ovaries of both are enclosed by bracteoles which are fused into a

cupule and are subtended by two bracts. The ovaries and bracteoles develop into

a single compound fruit. (b) Normal compound fruit produced by Haskap. A

single fruit results from each two-flowered inflorescence. (c) Haskap fruit in

which the bracteoles have not fused around the ovaries, showing the two distinct

berries which make up the fruit (Photo credit: Logie Cassells, LaHave Natural

Farms, Nova Scotia, Canada).

Table 1. The results of a negative binomial GLM comparing the number of pollen grains

found on the stigmas of Haskap (Lonicera caerulea L.) flowers after a single visit

by Apis mellifera L., Bombus spp. or Osmia lignaria Say to an unvisited control.

Estimate SE t value P-value 95% CI Intercept (Control) 4.28 0.08 52.72 < 0.001* 4.11 to 4.46 Apis mellifera 0.17 0.11 1.58 0.116 -0.07 to 0.39 Bombus spp. 0.32 0.11 2.91 0.004* 0.08 to 0.55 Osmia lignaria 0.40 0.21 1.85 0.065 -0.05 to 0.89 *indicates a significant effect (P<0.05)

Figure 2. Single visit pollen deposition (SVD) of pollen on Haskap (Lonicera caerulea

L.) stigmas by three taxa of bees. SVD is derived from the predicted estimates of

a generalized linear model, and represents the total amount of pollen found on

visited stigmas minus the pollen on unvisited controls, the result of internal self-

pollination. The dotted line represents the minimum pollination requirements

(i.e., 24 grains deposited) for each flower to ensure complete fertilization of all

the ovules (see Appendix B, Supplementary Material).

Figure 3. Mean (± SE) densities of Bombus spp. and Apis mellifera L. in a Haskap

(Lonicera caerulea L.) orchard at each hour from dawn until dusk during

flowering. The data are presented as number of individuals per 400 open flowers.

Figure 4. The relationship between density of foragers and air temperature (˚C) for Apis

mellifera and Bombus spp. across 99 surveys performed from 24-May to 10-June-

2014 in a 30ha organic Haskap orchard. Surveys were performed by scan

sampling, and a total of 400 flowers were observed per survey.

Table 2. The results of a negative binomial GLM regression on the response of Apis

mellifera L. and Bombus spp. densities to the variables time, temperature, wind,

cloud cover and density of the other taxon.

Taxon Variable Estimate SE t-value P-value 95% CI Density of Intercept -1.89 0.80 -2.37 0.0176 -4.16 to 0.52 foraging A. Time -0.07 0.04 -1.64 0.1021 -0.18 to 0.04 mellifera Temperature 0.26 0.04 6.67 >0.001* 0.16 to 0.38 Wind 0.02 0.03 0.78 0.4382 -0.04 to 0.08 Cloud cover 0.00 0.00 0.21 0.8354 -0.01 to 0.01 Bombus spp. density 0.08 0.05 1.53 0.1261 -0.02 to 0.18 Density of Intercept -0.41 0.46 -0.90 0.3682 -1.37 to 0.53 foraging Time 0.09 0.02 3.74 >0.001* 0.04 to 0.13 Bombus spp. Temperature -0.01 0.02 1.74 0.0822 -0.01 to 0.09 Wind 0.00 0.01 -0.81 0.4167 -0.04 to 0.02 Cloud cover 0.00 0.00 -1.64 0.1014 -0.01 to 0.00 A. mellifera density 0.01 0.00 1.48 0.1389 0.00 to 0.01 *indicates a significant effect (P<0.05)

CHAPTER 3: Floral longevity, nectar production and pollen release in Haskap

(Lonicera caerulea L.: Caprifoliaceae)

1. Introduction

Haskap (Lonicera caerulea L.: Caprifoliaceae), also known as blue honeysuckle or honeyberry, is an early flowering, temperate fruiting shrub native to northern parts of

Europe, Asia and North America. Its tart blue berry-like fruit are suspected of having various health benefits (Svarcova et al. 2007), and it is growing in popularity as a commercial crop, particularly in North America. Existing research on Haskap cultivation has focused largely on the shape, taste, and harvestability of this fruit. However, despite

Haskap being self-incompatible (Bors 2008) and requiring insect pollinators in order to set fruit (Bożek 2012), there has been little focus on its pollination biology. Research on cultivars of Haskap grown in Poland have found that it produces abundant pollen and nectar that is favoured by honey bees and bumble bees, as well as a variety of solitary bees (Bożek & Wieniarska 2006; Bożek 2007). Similar pollinator guilds appear to visit

North American cultivars, and field studies have suggested that bumble bees may be one of the most important pollinators for this cold-adapted crop (Frier et al. 2016). However, there is still little known about the specific diversity of floral visitors and their relative value, including the potential for nocturnal pollinators like moths, which are known to be important for other Lonicera species (Miyake & Yahara 1998). Additionally, although floral structure and fruit development is unique (Fig. 1), little is known about specific floral traits or reproductive strategies of Haskap.

Nectar and pollen are the most common pollinator rewards produced by flowers, and many pollinators, especially bees, rely entirely on these resources in one or all of their life stages (Proctor 1996; Armbruster 2012). Therefore, the nutritive composition, abundance, and availability of these rewards, are often indicative of the pollinator guilds

38 associated with the flowers (Fenster & Armbruster 2004). This may be especially true of nectar, which is produced specifically to attract pollinators (unlike pollen, where the reward function is secondary), and its production appears to adapt more quickly to different pollinator guilds than other floral traits (Ackermann & Weigend 2006). Many species are found to have distinct patterns of floral nectar production, nutritive composition, and resorption rates that may be correlated to the abundance, behaviour, and physiology of their associated pollinators (Galetto & Bernardello 1993, 1996;

Galetto et al. 1997; Witt et al. 1999; Nepi et al. 2001; Fenster & Armbruster 2004; Wolff et al. 2006; Kulloli et al. 2011; Agostini et al. 2011; Amorim et al. 2013). This reflects the concept of pollination syndromes, whereby a suite of phenotypic floral traits are thought to correlate to the primary functional pollinator guild, and contrasts with the theory that most plants are in fact generalists that are visited by a wide variety of pollinator guilds (i.e., poliphily) (Waser et al. 1996; Ollerton & Watts 2000; Freitas &

Sazima 2006; Johnson & Nicolson 2008; Petanidou et al. 2008; Ollerton et al. 2009). In either case, understanding the patterns of nectar production and other intra-floral aspects during anthesis may be especially important for agriculturally significant plants, where growers must decide whether to introduce managed pollinators, and more specifically, what species will be most effective. Currently, the nectar dynamics of Haskap are entirely unknown; information on nectar production as well as other important aspects of anthesis will help refine and optimize the pollination strategy for this crop and ultimately help growers to maximize commercial yield.

The purpose of our study was to determine floral longevity, nectar dynamics, pollen release, and stigma receptivity of Haskap flowers to learn more about the plant’s

39 reproductive strategy and its associated animal pollinators. Our specific objectives were to determine: (1) how long flowers are open, and whether floral longevity is affected by pollination; (2) when nectar is produced during anthesis, and whether total nectar production is affected by nectar removal; (3) when pollen is released from the anthers; and (4) how early in anthesis the stigma becomes receptive.

2. Methods

2.1 Study site and plants

All experiments were performed during February 2015 at the University of

Saskatchewan (Saskatoon, Saskatchewan), in a glass research greenhouse with daytime temperatures held at 15˚C and nighttime temperatures at 10˚C. The Haskap plants were grown under full spectrum lighting between 6AM and 11 PM, with supplementation by natural light. We used a total of 8 Haskap bushes from the cultivar ‘Tundra’ for our experimental treatments, and pollen from a Japanese cultivar compatible with Tundra was used as a pollen source (we hereafter refer to this cultivar as the pollinizer). Each potted plant was seven years old (maintained by the breeding program at the University of Saskatchewan), and kept in cold storage for a period of winter dormancy. Upon introduction to the greenhouse, the plants required approximately 10-14 days to begin flowering. Plants were watered every 2 days. As other researchers were using this space, a commercially purchased colony of Bombus impatiens Cresson (Hymenoptera: Apidae) was introduced to the greenhouse for pollination four days before the completion of our experiments. To prevent visitation in our experiments, flowers that remained to be

40 sampled when the bees were present were covered with pollinator exclusion bags

(Delnet® Pollination™ Bags).

2.2 Length of anthesis

Four experimental treatments were used to determine the duration of flowering in an inflorescence and the effects of pollination on anthesis length: (1) anthers removed before opening, stigma not pollinated; (2) anthers left intact, stigma not pollinated; (3) anthers left intact, stigma hand-pollinated; and (4) anthers removed, stigma hand- pollinated. In each case, the treatments were applied to both flowers in the inflorescence.

Each potted plant received 3 replicates of each treatment, and the treatments were assigned in a randomized order as inflorescences opened. Anthers were removed immediately after flower opening, before they had dehisced, and hand pollination was performed in the morning, after the inflorescences had been open for at least 24 hours.

Freshly collected pollen from a nearby pollinizer was applied to the stigmas of each inflorescence by hand. The inflorescences were observed once in the morning (9 AM), afternoon (2 PM) and evening (7 PM), and the beginning and end of anthesis were recorded, with the end marked by separation of the corollas of both flowers from the ovaries.

The duration (h) of anthesis for each treatment was compared using a Kruskal-

Wallis test, followed by multiple comparisons using the Mann-Whitney U test, with p- values adjusted using the Bonferonni correction. In addition, we calculated the percentage of flowers (including flowers used in nectar and pollen analysis) that were

41 first observed to be open in the morning, versus the afternoon or evening. All statistical analysis was performed in R (R Core Team 2014).

2.3 Nectar, pollen and stigma dynamics

In order to determine nectar and pollen dynamics, we measured nectar volume

(pooled from both flowers in the inflorescence), and recorded the number of open anthers per inflorescence at 9 separate intervals (3 per day) throughout anthesis. Upon opening, an inflorescence was randomly assigned an interval between 1 and 9; this was replicated three times per bush. Intervals always began at 9AM; if an inflorescence opened in the morning, the interval 1 began immediately, and if it opened in the afternoon or evening, interval 1 began the following morning. Each inflorescence was only sampled once. Nectar sampling was performed by inserting a 5µL micro-capillary tube into the bottom of the corolla to draw up the nectar. The total nectar volume (Vn) was calculated by measuring how much of the tube was filled, and then applying the

5휇푙 formula 푉푛 = ∗ 퐿푛. Lc is the length of the capillary tube until the 5µL mark, and Ln is 퐿푐 the length of the tube filled by the nectar sample.

Nectar production and pollen release were analyzed first by interval, to assess how they varied by time of day. However, because inflorescences did not always open in the morning, the actual time an inflorescence had been open at each interval varied. To account for this, we then calculated the number of hours each inflorescence had been open at time of sampling to produce nine intervals of 8 hours each to examine nectar production and pollen release over hours of anthesis. Both scenarios were analyzed using a Kruskal-Wallis test with the sampling interval as the predictor variable and nectar

42 volume as the response, followed by multiple comparisons using the Mann-Whitney U test, with the P-values adjusted using the Bonferroni correction.

To test for the effect of nectar removal on total nectar production, three inflorescences were randomly selected from each bush. Starting at the beginning of anthesis, we removed the nectar from each of the inflorescences at 7PM each day after opening, and recorded the amount of nectar. This was repeated daily until the end of anthesis.

The effect of nectar removal on total nectar production was analyzed using the same method, with sampling day as the predictor and nectar volume as the response. We then added the total amount of nectar produced by each re-sampled inflorescence, using only inflorescences that were sampled 3 times total, to get the average nectar produced over the lifetime of the inflorescence. This was compared to the average maximum volume of nectar produced by single-sampled inflorescences using a Mann-Whitney U test.

To determine when the stigmas become receptive with respect to stage of anthesis, we removed the stigmas from 3 inflorescences at different stages of opening, beginning with inflorescences that were still tightly closed. The stigmas were removed from each flower using forceps, and the styles were clipped towards the base as far down as possible. We ensured there were no pollen grains on the stigma using a hand lens, and then held the stigma in a 3% hydrogen peroxide solution and watched for the formation of bubbles on the stigma surface, which would indicate the presence of peroxidases

(Dafni et al. 2005). This was repeated at progressively more advanced stages of anthesis until the stigmas showed receptivity. As old stigmas can react with hydrogen peroxide

43 even though they are no longer receptive (Dafni & Maues 1998), only the beginning of receptivity was identified.

3. Results

3.1 Length of anthesis

Though anthesis began at various times throughout the day, the majority [i.e.,

60.7% (162/267)] were first observed to be open in the morning, versus 18.3% in the afternoon, and 21.0% in the evening. Anthesis length differed between treatments

(H=26.9 df=3, P=6.1-06; Fig. 2), with pollinated inflorescences flowering for a shorter duration [59.5± 16.5 h (mean ± SD)] than un-pollinated inflorescences (83.4 ± 25.9 h).

There was no effect of emasculation on length of anthesis. Following hand-pollination, anthesis lasted another 34.3 ± 15.2 hours.

3.2 Nectar, pollen and stigma dynamics

Nectar was evident as soon as the individual flowers opened and may be produced even before the onset of anthesis. Nectar volume increased during the first morning and afternoon, and stabilized by the end of the first full day, with maximum average volume occurring the second morning (Kruskal-Wallis test, Χ2 = 28.8, df = 5, P

< 0.001; Fig. 3). Discrete hourly intervals showed similar results; nectar production began immediately and volume did not change significantly after 16 hours of anthesis, although maximum average volume was observed at hour 40 (Kruskal-Wallis test, Χ2 =

41.2, df = 8, P < 0.001; Fig. 4). In both cases, nectar volume was maintained throughout

44 anthesis and nectar was not resorbed. We observed that the corollas of the flowers would often abscise from the ovaries with nectar still present.

Anthers began to dehisce almost immediately after flowers opened, and proceeded rapidly over the first day, with the majority of pollen presented by the second afternoon (Kruskal-Wallis test, Χ2= 92.7, df = 8, P < 0.001; Fig. 3). Analysis by discrete hourly intervals suggests pollen is released almost invariably within the first 24 hours

(Kruskal-Wallis test, Χ2 = 130.2, df = 8, P < 0.001; Fig. 4).

After removal, nectar was wholly replenished throughout anthesis (Fig. 5).

Maximum nectar volume recorded at the end of the second day was significantly higher than on day 1, supporting our previous finding that maximum nectar production occurs during the second day. Total nectar produced by flowers with their entire volume twice removed was significantly higher than those without nectar removal (6.1 ± 3.8 µL compared to 2.3 ± 1.7 µL; Mann-Whitney U test, W = 27, P < 0.001)

Evidence of peroxidase activity on the stigmas was observed in flowers that were still tightly closed, suggesting that the stigma may be receptive even before the onset of anthesis.

4. Discussion

The timing of commencement of anthesis is staggered; flowers open throughout the day, and likely during the night as well. The majority of flowers were first observed to be open at 9AM; since observations were not made overnight, it’s possible that many of the flowers observed at this time actually opened many hours earlier, as anthesis appeared quite advanced in some of these flowers. This could indicate that nocturnal

45 insects such as moths may be important contributors to Haskap pollination, in addition to diurnal bees and flies. The average duration of anthesis of the flowers is about 3 to 4 days (Fig. 2), however, pollination will significantly shorten this, triggering senescence of the flowers to within about a day. Our findings are slightly shorter than those of

Bożek and Wieniarska (2006) who found that flowers of Lonicera caerulea var. kamtschatica lived 4-5 days, but they also noted that flowers excluded from pollinators were longer lived. This extended floral period means more opportunities for visitation by insects and increased pollination success, while senescence in response to pollination reduces the occurrence of repeat visits to flowers that are already pollinated, and may also reduce damage to the flowers, which can reduce seed set (Young 1988; Burquez &

Corbet 1991). The fact that emasculation has no effect on floral longevity suggests the plants are completely self-incompatible, and self-pollination does not shorten the lifespan of the flower, nor does it result in significant stigma clogging that may prevent cross pollination.

However, we did not quantify actual levels of self-pollination in this study, but our field studies have shown significant self-pollination of the stigma (Frier et al. 2016).

It is possible that in greenhouse conditions, very little self-pollination actually occurs, and in more realistic settings (with vigorous disruption by wind and handling by insect visitors), these factors may be more significant. This is made more likely by the fact that pollen is released almost immediately following the onset of anthesis (Fig 3 and Fig 4), leaving little or no prior opportunity for cross-pollination. This strategy could maximize the chance that pollen is picked up by a floral visitor, but it also could increase the chance of self-pollen interfering with cross-pollination of the stigma and reducing

46 reproductive success (Bertin & Sullivan 1988; Galen et al. 1989; Waser & Price 1991;

Broyles & Wyatt 1993; Barrett 2002), especially because the anthers and stigma exist in very close proximity to one another. If this type of stigma clogging is common, this could suggest some competition or trade-off between male and female reproductive success in Haskap. It may be worthwhile to explore differences in style length among

Haskap cultivars, as well as compared to wild varieties, as longer styles may be less likely to experience self-pollination.

Our analysis of nectar production by intervals organized by time of day (Fig. 3) is congruent to those organized by discrete intervals of time after opening (Fig. 4). Due to the lack of overlap in the age of flowers in the latter, we base our conclusions on this analysis, and we have no evidence to suggest that nectar production is affected by circadian rhythms. Nectar production begins as soon as the flowers open, perhaps even slightly before, during the bud stage. We have observed bumble bees visiting Haskap flowers before they are entirely open, and our results suggest that the stigma is receptive at this stage as well. Early nectar production and stigma receptivity may have evolved to take advantage of these early visitors and increase the chance of successful pollination.

Nectar volume does not change significantly after the first 16 hours and is maintained throughout the remainder of anthesis. Considerable variation in the amount of nectar is consistent with findings that variability in nectar production is correlated with large floral displays (Biernaskie & Cartar 2004), as risk-averse pollinators pay shorter visits to a single bush when nectar is variable (Biernaskie et al. 2002). As Haskap produces many flowers simultaneously and is self-incompatible, this strategy would help decrease the instance of geitonogamy and promote cross-pollination.

We found no evidence of net nectar resorption, and the corolla abscised from the ovaries with the nectar load intact. However, we only analyzed nectar dynamics in un- pollinated flowers; in some flowers, reabsorption is triggered by pollination, presumably to reuse the energy resources in berry development (Luyt & Johnson 2002). However,

Burquez and Corbet (1991) suggest that if the nectary is lost when the corolla dehisces, as in Lonicera, reabsorption is unlikely. Additionally, the entire nectar volume can be replaced several times throughout anthesis. This may be an adaptation to nectar robbers or ineffective pollinators, increasing the possibility of repeat visits to a single flower. We have commonly observed both honey bees and bumble bees nectar robbing the flowers, and there is evidence that many legitimate pollinator visits do not deposit sufficient pollen grains for full fertilization of the ovaries (Frier et al. 2016).

5. Conclusion

The results of this study suggest that Haskap flowers are likely generalized in their pollinator attraction strategy and may be pollinated by a wide variety of insect species and functional groups. Haskap flowering occurs very early in the year when few pollinators are active, and the characteristics described here likely reflect adaptations to capitalize on every opportunity to receive a successful pollination visit. The flowers, which open throughout the day (and probably the night as well), remain open for up to four days, but successful pollination triggers early senescence. Nectar is produced immediately upon anthesis, potentially beginning in the large bud stage, and after initial nectar production, the volume is held relatively constant until senescence and any removed nectar during this period is replaced. The anthers begin to dehisce immediately

48 after the flower opens and continue over the first day. The stigma appears receptive in the bud stage, and potentially could be pollinated before the flower opens.

As a plant with polyphilic flowers, Haskap is less likely to be pollinator limited than more specialized species, and it may be effectively pollinated by a wide variety of managed and wild insects. This means that Haskap could be successfully cultivated in a wide variety of habitats and geographic locations, and not be limited geographically or to commercial use of specific pollinators. As the flowers open throughout the day and into the night, it is also likely that nocturnal insects such as moths are important pollinators of this crop, not just diurnal bees and flies. Haskap growers should take advantage of this generalist system by developing and maintaining healthy pollinator habitat in and around

Haskap orchards, as species-rich pollinator populations may be essential to realizing optimum fruit yields from this crop (Frier et al. 2016)

References

Ackermann M, Weigend M (2006) Nectar, floral morphology and pollination syndrome in Loasaceae subfam. Loasoideae (Cornales). Annals of Botany 98:503–514. Agostini K, Sazima M, Galetto L (2011) Nectar production dynamics and sugar composition in two Mucuna species (Leguminosae, Faboideae) with different specialized pollinators. Die Naturwissenschaften 98:933–42. Amorim FW, Galetto L, Sazima M (2013) Beyond the pollination syndrome : nectar ecology and the role of diurnal and nocturnal pollinators in the reproductive success of Inga sessilis (Fabaceae). Plant Biology 15:317–327. Armbruster WS (2012) Evolution and ecological implications of “specialized” pollinator rewards. In: Patiny S (ed) Evolution of Plant-Pollinator Relationships. Cambridge University Press, pp 44–67. Barrett SCH (2002) Sexual interference of the floral kind. Heredity 88:154–159. Bertin RI, Sullivan M (1988) Pollen interference and cryptic self-fertility in Campsis

radicans. American Journal of Botany 75:1140–1147. Biernaskie JM, Cartar RV (2004) Variation in rate of nectar production depends on floral size: a pollinator hypothesis display manipulation hypothesis. Functional Ecology 18:125–129. Biernaskie JM, Cartar RV, Hurly TA (2002) Risk-averse inflorescence departure in hummingbirds and bumble bees: could plants benefit from variable nectar volumes? Oikos 98:98–104. Bors B (2008) Growing Haskap in Canada. U of S Fruit Program Bożek M (2007) Pollen productivity and morphology of pollen grains in two cultivars of honeyberry (Lonicera kamtschatica (Sevast.) Pojark.). Acta Agrobotanica 60:73– 77. Bożek M (2012) The effect of pollinating insects on fruiting of two cultivars of Lonicera caerulea L. Journal of Apicultural Science 56:5–11. Bożek M, Wieniarska J (2006) Biologia kwitnienia i wydajność cukrowa kwiatów dwóch odmian Lonicera kamtschatica (Sevast.) Pojark / Blooming biology and sugar efficiency of two cultivars of Lonicera kamtschatica (Sevast.) Pojark. Acta Agrobotanica 59:177–182. Broyles S, Wyatt R (1993) The consequences of self-pollination in Asclepias exaltata, a self-incompatible milkweed. American Journal of Botany 80:41–44. Burquez A, Corbet SA (1991) Do flowers reabsorb nectar? Functional Ecology 5:369– 379. Dafni A, Kevan PG, Husband BC (2005) Practical Pollination Biology. Enviroquest Ltd., Cambridge, Ontario. Dafni A, Maues MM (1998) A rapid and simple procedure to determine stigma receptivity. Sexual Plant Reproduction 11:177–180. Fenster C, Armbruster W (2004) Pollination syndromes and floral specialization. Annual Review of Ecology, Evolution, and Systematics 35:375–403. Freitas L, Sazima M (2006) Pollination biology in a tropical high-altitude grassland in Brazil: interactions at the community level. Annals of the Missouri Botanical Garden 93:465–516. Frier SD, Somers CM, Sheffield CS (2016) Comparing the performance of native and managed pollinators of Haskap (Lonicera caerulea: Caprifoliaceae), an emerging fruit crop. Agriculture Ecosystems and Environment 219, 42-48.

Galen C, Gregory T, Galloway L (1989) Costs of self-pollination in a self-incompatible plant, Polemonium viscosum. American Journal of Botany 76:1675–1680. Galetto L, Bernardello L (1993) Nectar secretion pattern and removal effects in three species of Solanaceae. Canadian Journal of Botany 71:1394–1398. Galetto L, Bernardello G (1996) Characteristics of nectar secretion by Lycium cestroides, L. ciliatum (Solanaceae), and their hybrid. Plant Species Biology 11:157–163. Galetto L, Bernardello G, Rivera GL (1997) Nectar, nectaries, flower visitors, and breeding system in five terrestrial Orchidaceae from central Argentina. Journal of Plant Research 110:393–403. Johnson SD, Nicolson SW (2008) Evolutionary associations between nectar properties and specificity in bird pollination systems. Biology Letters 4:49–52. Kulloli SK, Chandore AN, Aitawade MM (2011) Nectar dynamics and pollination studies in three species of Lamiaceae. Current Science 100:509–516. Luyt R, Johnson S (2002) Postpollination nectar reabsorption and its implications for fruit quality in an epiphytic orchid. Biotropica 34:442–446. Miyake T, Yahara T (1998) Why does the flower of Lonicera japonica open at dusk? Canadian Journal of Botany 76:1806–1811. Nepi M, Guarnieri M, Pacini E (2001) Nectar secretion , reabsorption , and sugar composition in male and female flowers of Cucurbita pepo. International Journal of Plant Sciences 162:353–358. Ollerton J, Alarcón R, Waser NM, Price MV., Watts S, Cranmer L, Hingston A, Peter CI, Rotenberry J (2009) A global test of the pollination syndrome hypothesis. Annals of Botany 103:1471–1480. Ollerton J, Watts S (2000) Phenotype space and floral typology: towards an objective assessment of pollination syndromes. Det Norske Videnskaps-Akademi. I. Matematisk Naturvidenskaplige Klasse, Skrifter, Ny Serie 39:149–159. Petanidou T, Kallimanis AS, Tzanopoulos J, Sgardelis SP, Pantis JD (2008) Long-term observation of a pollination network: Fluctuation in species and interactions, relative invariance of network structure and implications for estimates of specialization. Ecology Letters 11:564–575. Proctor M (1996) The Natural History of Pollination. Whitecap Books. R Core Team (2014) R: A language and environment for statistical computing. Svarcova I, Heinrich J, Valentova K (2007) Berry fruits as a source of biologically active

compounds: the case of Lonicera caerulea. Biomedical Papers 151:163–174. Waser NM, Chittka L, Price MV., Williams NM, Ollerton J (1996) Generalization in pollination systems, and why it matters. Ecology 77:1043–1060. Waser N, Price M (1991) Reproductive costs of self-pollination in Ipomopsis aggregata (Polemoniaeae): are ovules usurped? American Journal of Botany 78:1036–1043. Witt T, Geyer R, Gottsberger G (1999) Nectar dynamics and sugar composition in flowers of Silene and Saponaria species (Caryophyllaceae). Plant Biology 1:334– 345. Wolff D, Witt T, Ju A, Gottsberger G (2006) Nectar dynamics and reproductive success in Saponaria officinalis (Caryophyllaceae) in southern Germany. Flora 201:353– 364. Young HJ (1988) Differential importance of beetle species pollinating Dieffenbachia longispatha (Araceae). Ecology 69:832–844.

Figure 1. A typical Haskap inflorescence. The downward facing flowers are pale yellow

and each has five petals, five stamens and one stigma; the ovaries of both flowers

are enclosed by the bracteoles, which have fused into a cupule. The ovaries and

bracteoles develop into the Haskap ‘berry’, actually a multiple fruit.

Figure 2. The length of anthesis (h), from opening of the bud to abscission of the

corolla, for un-pollinated and pollinated Haskap inflorescences that had the

anthers removed or left intact. Vertical bars represent the 95% confidence

interval. Bars with different letters are significantly different at P < 0.05

according to pairwise comparisons using a Mann-Whitney U test with p-values

adjusted for multiple comparisons using the Bonferroni correction. Sample size is

reported in brackets beneath each bar.

Figure 3. The nectar volume (µL) (left y-axis) and the number of open anthers (right y-

Axis) in Haskap (Lonicera caerulea) inflorescences per time interval (x-axis),

starting the first morning the flowers were open until the end of flowering (i.e.,

inflorescences that opened after 9am were begun sampling the following

morning). Vertical bars represent the 95% confidence interval. Plots with

different letters are significantly different at P < 0.05 according to pairwise

comparisons using a Mann-Whitney U test with p-values adjusted for multiple

comparisons using the Bonferroni correction. Sample size is indicated in brackets

under each treatment.

Figure 4. The amount of nectar (µL) (left y-axis) and the number of open anthers (right

y-axis) in Haskap (Lonicera caerulea) inflorescences per 8 hour interval after the

onset of anthesis (x-axis). Vertical bars represent the 95% confidence interval.

Plots with different letters are significantly different at P < 0.05 according to

pairwise comparisons using a Mann-Whitney U test with p-values adjusted for

multiple comparisons using the Bonferroni correction. Sample size is indicated in

brackets under each treatment.

Figure 5. The amount of nectar (µL) in Haskap (Lonicera caerulea) inflorescences that

had nectar removed once per day, up to 3 consecutive days. Vertical bars

represent the 95% confidence interval. Plots with different letters are

significantly different at P < 0.05 according to pairwise comparisons using a

Mann-Whitney U test with p-values adjusted for multiple comparisons using the

Bonferroni correction. Sample size is indicated in brackets under each treatment.

CHAPTER 4: General Discussion

4.1 Synthesis

Despite a long history of cultivation in Europe and Asia, Haskap (Lonicera caerulea L.) is relatively unknown within North America. However, since the establishment of a Haskap breeding program at the University of Saskatchewan (Bors et al., 2012) which has released several new cultivars, small scale production of the crop has begun in most Canadian provinces and the Yukon Territory (University of

Saskatchewan Fruit Program, n.d.). The flavourful blue fruit are being marketed as the newest “superfood” due to their nutraceutical properties and perceived health benefits

(Chaovanalikit et al., 2004; Raudsepp et al., 2013; Rop et al., 2011; Rupasinghe et al.,

2012; Svarcova et al., 2007), and the fruits are eaten fresh or used in a variety of food products such as jam, ice cream, and wine (Bors, 2008).

Haskap flowers are self-incompatible and require insect pollinators in order to set fruit (Bors et al., 2012; Hummer et al., 2012). However, very little has been studied regarding the pollination and reproductive biology of this crop. Currently, the honey bee

(Apis mellifera L.) is the main managed pollinator used for this crop, but this species may not be climatically suited for pollinating this early season crop which flowers when cool temperatures may discourage workers foraging (Hummer et al., 2012; Tan et al.,

2012). As a result, there is some interest in using other commercial pollinators, including blue orchard bees (Osmia lignaria Say.), as an alternative managed pollinator. In addition, as the crop likely receives substantial pollination services from wild pollinators such as bumble bees (Bombus spp.) (Hummer et al., 2012), there is also interest in ensuring that their populations are conserved. My research aimed to compare the performance of these three taxa as pollinators of this crop, in order to assess the potential

59 value of different pollination management strategies. In addition, I described the length of anthesis, nectar and pollen dynamics, and stigma receptivity of Haskap flowers, which are unique in terms of structure and fruit development.

In Chapter 2, the performance of the three target taxa, A. mellifera, O. lignaria, and Bombus spp., was compared as it pertains to Haskap pollination. We observed that unvisited flowers had high levels of pollen on the stigma (presumably due to intra-floral pollen transfer), which made it difficult to determine exact levels of inter-floral pollination by each taxon and has important implications for future studies of Haskap pollination, which we discuss. However, we were able to determine that Bombus spp. were the most effective and efficient pollinators of Haskap; they had the highest levels of single visit pollen deposition (SVD), the highest rate of floral visitation, and actively foraged even in low temperatures. Apis mellifera workers performed poorly as individuals; they had low SVD, low visit rate, and were less likely than the other taxa to visit both flowers in an inflorescence, which was shown to be an important factor for fruit development (Chapter 2, Appendix B). Despite this, as high numbers of workers are easily available in managed colonies, they exhibited much higher densities than Bombus spp., and may contribute significantly to overall fruit production through “brute force” pollination (Morse, 1991). However, it is important to note that, unlike Bombus, they did not forage during lower temperatures, which are commonly observed during Haskap flowering. In contrast to A. mellifera, O. lignaria females performed well in measures of

SVD, visitation rate, and visit symmetry, but analysis of their pollen loads indicated that they have a low affinity for Haskap, and prefer alternative forage such as Salix spp.

Chapter 3 describes anthesis length, nectar production timing and volume, timing of pollen release, and stigma receptivity period of Haskap cultivar ‘Tundra’, which was brought to bloom in a greenhouse at the University of Saskatchewan (Saskatoon,

Saskatchewan) campus. We found that the average flowering duration for un-pollinated inflorescences was between 3 and 4 days, and that pollination would trigger early senescence, within about a day of the pollination event. Emasculation of the flower had no significant effect on flowering duration, suggesting that the flowers are completely self-incompatible and that stigma clogging induced by autogamy does not affect pollination under these conditions. Nectar was produced very early in anthesis, potentially beginning in the bud stage. Following the first 16 hours of anthesis, nectar volume did not change significantly, and nectar removal at any point in anthesis triggered additional nectar production. Pollen release began at the start of anthesis, and all anthers had typically dehisced by the end of the first day. Stigmas were reactive to hydrogen peroxide even while still in the bud stage, suggesting that receptivity to pollen begins even before anthesis begins. Together, these characteristics suggest that Haskap flowers exhibit a generalist pollination strategy: long anthesis duration, early, abundant and replaceable nectar production, early pollen release, and early stigma receptivity, suggesting a strategy to maximize the probability of a visit in pollinator limited environments.

4.2 Conclusions and recommendations

Our results are consistent with many other studies (e.g. Artz and Nault, 2011;

Cane and Schiffhauer, 2003; Canto-Aguilar and Parra-Tabla, 2000; Garibaldi et al.,

2011, 2013; Greenleaf and Kremen, 2006; Javorek et al., 2002; Klein et al., 2003; Park et al., 2015; Rader et al., 2013; Rogers et al., 2013; Woodcock et al., 2013) that show that wild bees are important contributors to crop pollination. Although wild Bombus queens were present in much lower densities than A. mellifera workers, their higher effectiveness and efficiency as pollinators could compensate for this difference. We suggest that pollination strategies for Haskap emphasize the contributions of wild bees, and encourage the promotion of these important pollinators by providing natural areas with abundant forage and nesting habitat in areas surrounding Haskap orchards. Since individual Bombus that forage on Haskap are most often likely to be solitary queens in the nest initiation stage of their colony cycle, the number observed in a site is related to the size, health and productivity of colonies the previous fall (i.e., when the new queens are produced). Therefore, year-round forage for colony development is especially important for bumble bee colonies, and Haskap should be grown alongside crops that flower at later times, native wildflower meadows, or other natural habitat, as well as by allowing the growth of bee-friendly plants and ‘weeds’ such as dandelion. This will also help to promote other wild bees, which were not analyzed in our study but may still be important pollinators of Haskap. This is supported by the results of Chapter 3, which suggest that Haskap has a generalist-pollinator strategy and could be pollinated by a wide variety of organisms, including nocturnal insects. As well, high average SVD by O. lignaria suggests that other wild Osmia spp., who also carry their pollen loads in scopa located on the underside of their abdomen, could be important pollinators of Haskap.

Where local habitat is of poor quality for bees, or where Haskap orchards are very large, wild pollinators may not be abundant enough to facilitate high yields. In these

62 cases, A. mellifera colonies could be provided to enhance pollination. Apis mellifera workers will readily forage on Haskap flowers, and despite poor individual performance, the sheer number of foragers available in a managed colony means that they may be very effective in improving yields. However, it’s important to emphasize that the foraging activity of A. mellifera workers is dependent upon temperature, and cold spells during

Haskap flowering could leave much of the crop un-pollinated if wild Bombus queens are not present during these periods. Importantly, proximity to rich pollinator habitat is known to enhance pollination by A. mellifera as well as by wild bees (Decourtye et al.,

2010; Woodcock et al., 2013), and so habitat enhancement for pollinators should be a priority even if A. mellifera are expected to be the primary pollinators. In addition, it may be worth exploring the benefit of Haskap to apiaries; as Haskap flowers at a time when there is little other forage available, the flowers may be a good way to nourish A. mellifera colonies early in the year, and relationships between Haskap growers and apiarists may benefit both parties.

Although O. lignaria have many attributes that make them seem suitable as a

Haskap pollinator (i.e., their early emergence and limited foraging period), they do not appear to be a good candidate as a managed Haskap pollinator. Analysis of pollen collected by females showed that they prefer alternative early-available pollen sources such as Salix or Taraxacum spp., and are unlikely to visit Haskap with enough frequency to have a significant impact on fruit set. However, we did observe high average levels of

SVD for O. lignaria, and this could have important implications for their use as a greenhouse pollinator of Haskap. Commercially available Bombus colonies are often used in greenhouse settings, but currently the only managed species in Canada is Bombus

63 impatiens. This species is not native to western Canada, and for this reason the use of

Bombus impatiens in greenhouses in Saskatchewan is not recommended due to the risk of escape and introduction of a non-native species, as well as spread of associated pathogens. Therefore, O. lignaria could be a potential alternative, since it is found throughout Canada and is available in most areas, and in the absence of other food plants, may visit Haskap flowers for pollen.

4.3 Suggestions for further research

There is still a lot that is unknown with respect to Haskap pollination, and the opportunities for further study are many and varied. While the following does not constitute a complete list, we have identified three major areas for future study into the pollination biology of Haskap: comprehensive identification of floral visitors, linking pollinator performance to fruit set, and aspects of floral morphology.

This thesis has identified a preliminary list of floral visitors to Haskap at our study site, summarized in Appendix A, but it is by no means comprehensive. Systematic surveys should be done to identify pollinators that are associated with this crop throughout its range; not just for cultivated crops, but also for wild populations. Floral visitors can vary tremendously over spatial and temporal scales (Williams et al., 2001), and also with respect to local crop management practices (Morandin and Winston, 2005).

For example, preliminary sampling at an orchard near Outlook, Saskatchewan indicated a much higher diversity and abundance of solitary bees than in Birch Hills,

Saskatchewan. We have also observed hummingbirds and hawkmoths occasionally visiting the flowers, and they may be important pollinators for later-blooming varieties.

This diversity is not altogether unexpected; as our results from Chapter 3 suggest,

Haskap is a very generalist plant, it is probably adaptable to a wide range of pollinator guilds and communities, including nocturnal pollinators.

The next step in assessing taxa as pollinators is to bridge the gap between Haskap pollination and fruit set; although we have established that Bombus are effective pollinators, this may not translate linearly into yield (Cane and Schiffhauer, 2003;

Motten et al., 1981; Olsen, 1997). Experiments such as cage trials, or relating of pollinator species density to fruit yield in different orchards, could help to link the activity of different species to actual realized fruit set, as well as to determine how many individual pollinators are needed to fully pollinate a Haskap orchard. One important limitation of our methods was the inability to control for factors such as outcrossing and incompatibility; we could not differentiate between compatible (allogamous, or outcrossed) vs incompatible (autogamous and/or geitonogamous, or self) Haskap pollen.

Therefore, we were unable to identify whether certain species were more likely to deposit compatible (i.e., outcrossed) pollen grains; by following a single visit by a pollinator through to fruit set, it may be possible to elucidate differences in cross- pollination.

One of the most intriguing things about Haskap is its unique floral structure. No other Lonicera sp. shares the characteristic fused bracteoles and multiple fruit that is seen in Haskap (Rehder, 1909), nor any other species, so far as we know. The function of this dual structure is not clear, though it must have evolved through some relationship with pollinators, or perhaps seed dispersal. One line of inquiry might address potential differences between the two flowers; quantifying variation between the paired flowers in

65 nectar or pollen production, timing of anthesis, or stigma receptivity may hint at an evolutionary function (Bateman and Rudall, 2006; Guitián and Navarro, 1996; Itagaki and Sakai, 2006). One aspect of floral morphology we were unable to address was that of nectar sugar concentration and composition; we know from some preliminary studies that Haskap nectar is around 60% sugar, but we were unable to analyze how this might vary throughout anthesis (Agostini et al., 2011; Kulloli et al., 2011; Wist and Davis,

2008). In addition, floral characteristics could vary between cultivars, especially those derived from different germplasm. For example, we have noted differences in the length of the style in relation to the filaments between cultivars; this could have an effect on intra-floral pollen transfer, which can interfere with cross pollination (Barrett, 2002;

Galen et al., 1989; Wilcock and Neiland, 2002). In addition, breeding programs can unintentionally alter important aspects of floral biology, including floral scents and rewards, which can negatively impact the attractiveness of a cultivar to potential pollinators (e.g. Abrol, 1995; Klatt et al., 2013; Rodriguez-Saona et al., 2011). If these traits are not regularly assessed, fruit production could suffer due to insufficient pollination.

References

Abrol, D.P., 1995. Energetics of nectar production in some almond cultivars as a predictor of floral choice by honeybees Apis cerana indica F. and Apis mellifera L. (Hymenoptera: Apidae). Proc. Indian Natl. Sci. Acad. Part B Biol. Sci. 61, 285– 289. Agostini, K., Sazima, M., Galetto, L., 2011. Nectar production dynamics and sugar composition in two Mucuna species (Leguminosae, Faboideae) with different specialized pollinators. Naturwissenschaften 98, 933–42. doi:10.1007/s00114-011- 0844-6

Artz, D.R., Nault, B. A, 2011. Performance of Apis mellifera, Bombus impatiens, and Peponapis pruinosa (Hymenoptera: Apidae) as pollinators of pumpkin. J. Econ. Entomol. 104, 1153–1161. doi:10.1603/EC10431 Barrett, S.C.H., 2002. Sexual interference of the floral kind. Heredity (Edinb). 88, 154– 159. Bateman, R.M., Rudall, P.J., 2006. Evolutionary and morphometric implications of morphological variation among flowers within an inflorescence: a case-study using European orchids. Ann. Bot. 98, 975–93. doi:10.1093/aob/mcl191 Bors, B., 2008. Growing Haskap in Canada [WWW Document]. U S Fruit Progr. URL http://www.fruit.usask.ca/articles/growinghaskapinCanada.pdf (accessed 9.21.13). Bors, B., Thomson, J., Sawchuk, E., Reimer, P., Sawatzky, R., Sander, T., Kaban, T., Gerbrandt, E., Dawson, J., 2012. Haskap berry breeding and production - final report. Cane, J., Schiffhauer, D., 2003. Dose-response relationships between pollination and fruiting refine pollinator comparisons for cranberry (Vaccinium macrocarpon). Am. J. Bot. 90, 1425–1432. Canto-Aguilar, M., Parra-Tabla, V., 2000. Importance of conserving alternative pollinators: assessing the pollination efficiency of the squash bee, Peponapis limitaris in Cucurbita moschata (Cucurbitaceae). J. Insect Conserv. 4, 203–210. Chaovanalikit, A., Thompson, M.M., Wrolstad, R.E., 2004. Characterization and quantification of anthocyanins and polyphenolics in blue honeysuckle (Lonicera caerulea L.). J. Agric. Food Chem. 52, 848–852. doi:10.1021/jf030509o Decourtye, A., Mader, E., Desneux, N., 2010. Landscape enhancement of floral resources for honey bees in agro-ecosystems. Apidologie 41, 264–277. doi:10.1051/apido/2010024 Galen, C., Gregory, T., Galloway, L., 1989. Costs of self-pollination in a self- incompatible plant, Polemonium viscosum. Am. J. Bot. 76, 1675–1680. Garibaldi, L.A., Steffan-Dewenter, I., Kremen, C., Morales, J.M., Bommarco, R., Cunningham, S. S, Carvalheiro, L.G., Chacoff, N.P., Dudenhöffer, J.H., Greenleaf, S.S., Holzschuh, A., Isaacs, R., Krewenka, K., Mandelik, Y., Mayfield, M.M., Morandin, L. S, Potts, S.G., Ricketts, T.H., Szentgyörgyi, H., Viana, B.F., Westphal, C., Winfree, R., Klein, A.M., 2011. Stability of pollination services decreases with isolation from natural areas despite honey bee visits. Ecol. Lett. 14, 1062–72.

Garibaldi, L.A., Steffan-Dewenter, I., Winfree, R., 2013. Wild pollinators enhance fruit set of crops regardless of honey bee abundance. Science 339, 1608–1611. Greenleaf, S., Kremen, C., 2006. Wild bees enhance honey bees’ pollination of hybrid sunflower. PNAS 103, 13890–13895. Guitián, J., Navarro, L., 1996. Allocation of reproductive resources within inflorescences of Petrocoptis grandiflora (Caryophyllaceae). Can. J. Bot. 74, 1482–1486. Hummer, K.E., Pomper, K.W., Postman, J., Graham, C.J., Stover, E., Mercure, E.W., Aradhya, M., Crisosto, C.H., Ferguson, L., Thompson, M.M., Byers, P., Zee, F., 2012. Fruit Breeding. doi:10.1007/978-1-4419-0763-9 Itagaki, T., Sakai, S., 2006. Relationship between floral longevity and sex allocation among flowers within inflorescences in Aquilegia buergeriana var. oxysepala (Ranunculaceae). Am. J. Bot. 93, 1320–1327. Javorek, S.K., Mackenzie, K.E., Vander Kloet, S.P., 2002. Comparative pollination effectiveness among bees (Hymenoptera: Apoidea) on lowbush blueberry (Ericaceae: Vaccinium angustifolium). Ann. Entomol. Soc. Am. 95, 345–351. Klatt, B.K., Burmeister, C., Westphal, C., Tscharntke, T., von Fragstein, M., 2013. Flower volatiles, crop varieties and bee responses. PLoS One 8, e72724. doi:10.1371/journal.pone.0072724 Klein, A.-M., Steffan-Dewenter, I., Tscharntke, T., 2003. Fruit set of highland coffee increases with the diversity of pollinating bees. Proc. R. Soc. BB 270, 955–61. Kulloli, S.K., Chandore, A.N., Aitawade, M.M., 2011. Nectar dynamics and pollination studies in three species of Lamiaceae. Curr. Sci. 100, 509–516. Morandin, L.A., Winston, M.L., 2005. Wild bee abundance and seed production in conventional, organic, and genetically modified canola. Ecol. Appl. 15, 871–881. doi:10.1890/03-5271 Morse, R.A., 1991. Honeybees forever. Trends Ecol. Evol. 6, 337–338. doi:10.1016/0169-5347(91)90043-W Motten, A.F., Campbell, D.R., Alexander, D.E., 1981. Pollination effectiveness of specialist and generalist visitors to a North Carolina population of Claytonia virginica. Ecology 65, 1278–1287. doi:dx.doi.org/10.2307/1937292 Olsen, K., 1997. Pollination effectiveness and pollinator importance in a population of Heterotheca subaxillaris (Asteraceae). Oecologia 109, 114–121. Park, M.G., Raguso, R. A., Losey, J.E., Danforth, B.N., 2015. Per-visit pollinator

performance and regional importance of wild Bombus and Andrena (Melandrena) compared to the managed honey bee in New York apple orchards. Apidologie. doi:10.1007/s13592-015-0383-9 Rader, R., Reilly, J., Bartomeus, I., Winfree, R., 2013. Native bees buffer the negative impact of climate warming on honey bee pollination of watermelon crops. Glob. Chang. Biol. 19, 3103–3110. Raudsepp, P., Anton, D., Roasto, M., Meremäe, K., Pedastsaar, P., Mäesaar, M., Raal, A., Laikoja, K., Püssa, T., 2013. The antioxidative and antimicrobial properties of the blue honeysuckle (Lonicera caerulea L.), Siberian rhubarb (Rheum rhaponticum L.) and some other plants, compared to ascorbic acid and sodium nitrite. Food Control 31, 129–135. doi:10.1016/j.foodcont.2012.10.007 Rehder, A., 1909. Note on the morphology of the fruit of Lonicera caerulea. Rhodora 11, 209–211. Rodriguez-Saona, C., Parra, L., Quiroz, a., Isaacs, R., 2011. Variation in highbush blueberry floral volatile profiles as a function of pollination status, cultivar, time of day and flower part: implications for flower visitation by bees. Ann. Bot. 107, 1377–1390. doi:10.1093/aob/mcr077 Rogers, S.R., Tarpy, D.R., Burrack, H.J., 2013. Multiple criteria for evaluating pollinator performance in highbush blueberry (Ericales: Ericaceae) agroecosystems. Environ. Entomol. 42, 1201–9. doi:10.1603/EN12303 Rop, O., Řezníček, V., Mlček, J., Juríková, T., Balík, J., Sochor, J., Kramářová, D., 2011. Antioxidant and radical oxygen species scavenging activities of 12 cultivars of blue honeysuckle fruit. Hortic. Sci. 38, 63–70. Rupasinghe, H.P.V., Yu, L.J., Bhullar, K.S., Bors, B., 2012. Haskap (Lonicera caerulea): A new berry crop with high antioxidant capacity. Can. J. Plant Sci. 3, 1311–1317. doi:10.4141/CJPS2012-073 Svarcova, I., Heinrich, J., Valentova, K., 2007. Berry fruits as a source of biologically active compounds: the case of Lonicera caerulea. Biomed. Pap. 151, 163–174. Tan, K., Yang, S., Wang, Z.-W., Radloff, S.E., Oldroyd, B.P., 2012. Differences in foraging and broodnest temperature in the honey bees Apis cerana and A. mellifera. Apidologie 43, 618–623. doi:10.1007/s13592-012-0136-y University of Saskatchewan Fruit Program, n.d. Licensed Propagators [WWW Document]. URL http://www.fruit.usask.ca/propagators.html (accessed 10.27.15). Wilcock, C., Neiland, R., 2002. Pollination failure in plants: why it happens and when it

matters. Trends Plant Sci. 7, 270–7. Williams, N.M., Minckley, R.L., Silveira, F. a., 2001. Variation in native bee faunas and its implications for detecting community changes. Ecol. Soc. 5, 1–21. doi:7 Wist, T.J., Davis, A.R., 2008. Floral structure and dynamics of nectar production in Echinacea pallida var. angustifolia (Asteraceae). Int. J. Plant Sci. 169, 708–722. doi:10.1086/533602 Woodcock, B. A., Edwards, M., Redhead, J., Meek, W.R., Nuttall, P., Falk, S., Nowakowski, M., Pywell, R.F., 2013. Crop flower visitation by honeybees, bumblebees and solitary bees: Behavioural differences and diversity responses to landscape. Agric. Ecosyst. Environ. 171, 1–8. doi:10.1016/j.agee.2013.03.005

APPENDIX A – Floral visitors to Haskap

Table A.1 Observed floral visitors to Haskap (Lonicera caerulea L.) in an orchard near

Birch Hills, Saskatchewan in May, 2013. Individuals were identified through

several 30 minute surveys conducted over a three day period, as well as through

opportunistic observations and capture events. Bombus spp. were identified on

the wing whenever possible. Small solitary bees as well as unidentified Bombus

spp. were collected using a sweep net and were stored in ethanol to be identified

at a later time. Voucher specimens from this study are stored in the collection at

the Royal Saskatchewan Museum, Regina, Saskatchewan, Canada.

Order Family Species Apoidea Apidae Apis (Apis) mellifera Linnaeus, 1978 * Bombus (Pyrobombus) flavifrons Cresson, 1963 Bombus (Pyrobombus) huntii Greene, 1860 Bombus (Pyrobombus) melanopygus Nylander, 1848 Bombus (Pyrobombus) mixtus Cresson, 1878 Bombus (Bombias) nevadensis Cresson, 1874 Bombus (Pyrobombus) perplexus Cresson, 1863 Bombus (Pyrobombus) sylvicola Kirby, 1837 Bombus (Pyrobombus) ternarius Say, 1837 Bombus (Bombus) terricola Kirby 1837 Bombus (Pyrobombus) vagans Smith, 1854 Halictidae Halictus (Protohalictus) rubicundus Christ, 1791 Lasioglossum (Dialictus) laevissimum Smith, 1853 Lasioglossum (Dialictus) sagax Sandhouse, 1924 Lasioglossum (Dialictus) versans Lovell, 1905 Lasioglossum (Leuchalictus) zonulum Smith, 1848 Megachilidae Osmia (Osmia) lignaria Say, 1837 * Diptera Syrphidae Eristalis sp. *indicates an introduced and managed species

APPENDIX B – Pollination requirements for fruit set in Haskap

B.1 Methods

B.1.1 Pollination Treatments

To determine whether it is important for both flowers in the inflorescence to be pollinated, we assigned inflorescences to three treatment levels, where 1, 2, or 0 flowers were pollinated. Stigmas were hand pollinated using freshly collected pollen from a nearby pollinizer and applied using a wooden toothpick. We removed the stigmas of un- pollinated flowers to ensure they were not fertilized. Twenty Haskap bushes were used for this experiment; 10 from one row of ‘Tundra’ and 10 from one row of ‘Indigo Gem’.

Three branches were selected from each bush and randomly assigned to one of the 3 treatments. Each morning we observed the branches, and applied the appropriate treatment to flowers as they reached a stage just prior to anthesis. We repeated this daily until a total of 5 flowers had been treated per branch, resulting in a total of 100 inflorescences per treatment. Treatments began at the beginning of flowering, and it took

7 days to complete all treatments.

Just before harvest, fruit from each treatment were collected and categorized as either being a bud, semi-ripe (still red), or fully ripe. Haskap fruit in this orchard are harvested mechanically, which often results in the removal of non-ripe berries, therefore it was not possible to wait for all fruit to ripen. Percent fruit set was calculated per treatment by dividing the number of fruit obtained by the total number of inflorescences treated. The length and width of each fruit was measured using electronic calipers and

72 they were weighed using a compact digital scale, and then they were dissected to determine seed number.

The average length, width, weight, and seed number of fruits was averaged for each treatment per bush. We then used a generalized linear model (GLM) approach with appropriate error structures to compare each measured variable separately, with treatment and cultivar as predictor variables. Length, width and weight were all normally distributed, whereas seed number was fit to a negative binomial distribution. The relationship between weight and seed number was then determined using Pearson’s product-moment correlation.

B.1.2 Minimum pollen requirement for full fertilization

In order to determine the minimum amount of pollen required to fully fertilize each Haskap flower, we needed to know the maximum number of seeds possible within a berry, and average percent viability of the pollen. Pollen was collected from four cultivars within the orchard, including ‘Tundra’, ‘Indigo Gem’, ‘Borealis’, as well as the pollinizer. We collected freshly dehisced anthers from 100 inflorescences of each cultivar. For each cultivar, the anthers were mixed together and crushed to release the pollen, and 10 samples were taken and tested for viability using the MTT test for the presence of dehydrogenases in pollen (following Rodriguez-Riano, T., Dafni, A., 2000.

A new procedure to assess pollen viability. Sex. Plant Reprod. 12, 241–244). A minimum of 300 grains were counted in total from each sample, the average percent viability was calculated for each cultivar separately, and the cultivars were pooled for an overall average. The minimum pollen requirement for each flower was then calculated

73 by dividing the maximum number of seeds per fruit by 2 (obtained from the pollination treatment experiments) and dividing by the average percent viability of the pollen from all cultivars.

B.2 Results

Percent fruit set increased with increasing number of pollinated flowers in the inflorescence; 19/100 (19%) of the inflorescences set fruit when 0 flowers were pollinated, 64/100 (64%) when 1 flower was pollinated and 86/100 (86%) when 2 flowers were pollinated. Of these, 4, 49 and 59 fruit were ripe upon harvest, respectively; since only 4 fruit ripened from the treatment with 0 flowers treated, these were excluded from further analysis, though all four were considerably smaller than the other treatments.The mean of all measured variables was greater when 2 flowers were pollinated versus just 1 (Fig B.1), and significant differences were observed for weight, width, and number of seeds per fruit, though not for length. Cultivar was found to be a significant factor for fruit width, but not for length, weight or seed number (Table B.1).

Fruit weight and seed number, pooled between all three treatments, were positively correlated (r=0.73, t=11.05, df=110, p<0.001; Fig. B.2).

Average pollen viability for each cultivar ranged from 56 ± 3% (the pollinizer variety) to 72 ± 5% (‘Borealis’). ‘Tundra’ and ‘Indigo Gem’ had intermediate pollen viability of 61 ± 5% and 68 ± 5%, respectively. Overall, the average pollen viability was

64 ± 5%. The maximum number of seeds found in a fully pollinated fruit was 30 (15 per ovary), resulting in a per flower minimum pollen grain requirement of 24 grains total.

Figure B.1 The (a) length (mm), (b) width (mm), (c) weight (g) and (d) number of seeds

of Haskap (Lonicera caerulea L.) berries from two cultivars, ‘Indigo Gem’ and

‘Tundra’, in which one (‘Indigo Gem’, n=9; ‘Tundra’, n=9) or both (‘Indigo

Gem’, n=9; ‘Tundra’, n=8) of the flowers in the inflorescence were pollinated.

Each sample represents the average measurements from 1-5 berries per shrub,

and measurements were only taken from berries that were fully ripe at the time of

harvest. Lower, middle and upper quartiles are presented by the boxes and

maximum and minimum values by the vertical bars.

Table B.1 The results of generalized linear models comparing length, width, weight and

seed number in Haskap (Lonicera caerulea L.) fruit that had either one or both

flowers in the inflorescence pollinated.

Species Variable Estimate Std. Error z-value Pr(>|t|) 95% CI Length Intercept 15.53 7.76 2.00 0.0541 0.31 to 30.74 Treatment 0.44 0.28 1.55 0.1313 -0.12 to 0.99 Cultivar 0.00 0.01 -0.03 0.9784 -0.02 to 0.02 Width Intercept 21.81 4.40 4.96 <0.0001* 13.19 to 30.43 Treatment 0.48 0.16 3.01 0.0051* 0.17 to 0.79 Cultivar -0.01 0.00 -2.79 0.0088* -0.02 to 0.00

Weight Intercept 1.73 0.92 1.88 0.0696 -0.08 to 3.54 Treatment 0.11 0.03 3.21 0.0030* 0.04 to 0.17 Cultivar 0.00 0.00 -1.26 0.2183 0.00 to 0.00 Seed Intercept 5.27 2.66 1.98 0.0472* 0.07 to 10.48 Number Treatment 0.36 0.10 3.70 0.0002* 0.17 to 0.55 Cultivar 0.00 0.00 -1.58 0.1142 -0.01 to 0.00 * The mean difference is significant at the 0.05 level.

Figure B.2 Weight (g) vs seed number in Haskap fruit (r=0.73, t=11.05, df=106, p<

0.001).