Bumblebee worker parasites in the urban environment

Abstract
The aim of this project was to explore the ways in which urbanisation affects colonies of Bombus Terrestrius (Bumblebees) in the UK. This project examined parasite infection and ovarian development and found a falling trend in the rate of infection within infected colonies and the proportion of impervious surface. It also found a rising trend in the proportion of infection within a colony and the proportion of developed ovaries.
Glossary
Brood- Bee or wasp larvae.
Flagellum- A slender thread-like structure, especially a microscopic whip-like appendage which enables many protozoa, bacteria, spermatozoa, etc. to swim.
Graticule- A series of fine lines or fibres in the eyepiece of an optical device, such as a microscope, or on the screen of an oscilloscope, used as a measuring scale or an aid in locating objects.
Impervious- Not allowing fluid to pass through. For example, concrete.
Isotonic- A solution having the same osmotic pressure (salt concentration) as another solution, particularly one in a cell or body fluid.
Ovariole- A small ovary; specifically each of the structures in the ovary of an insect in which oocytes are formed, frequently tubular in shape.
Polygon- a shape with at least 3 straight sides and angles, typically five or more. For example, an irregular hexagon.
Protozoan- A Protozoan is a single-celled microscopic animal of a group of phyla of the kingdom Protista, such as an amoeba, flagellate, ciliate, or sporozoan.
All definitions were sourced from-
Oxforddictionaries.com. (2016). Oxford Dictionaries – Dictionary, Thesaurus, & Grammar. [online] Available at: http://www.oxforddictionaries.com/ [Accessed 11 Sep. 2016].
Introduction
Bumblebee populations in the UK have seen a decline in recent years. This has been attributed to changes in agricultural techniques such as increased use of pesticides⁵, and the destruction of hedgerow⁶ reducing the number of wild flowers available to bees. However, this project has focused on how urbanisation affects (if at all) infection through parasites in bumblebee populations. This was measured by the proportion of impervious surface in the area surrounding each bumblebee colony . This method has been used previously as a way of quantifying the effect urbanization on insects^{18, 21}.
When bees get inside of a home, they can cause real problems with walls not offering the ventilation needed to keep the honey from rotting. Despite the potential damages, bee extermination should never be done. Using a Honey Bee Relocation directory you can find a humane bee relocation service to come remove the hive and honeycomb from the structure without killing the honeybees.
Interestingly this project yielded a surprisingly low number of parasite positive samples with only around 10 percent of samples having parasites of any kind. This can be largely attributed to seasonal variations which can cause huge variations in parasite transmission. In a study by Shykoff and Schmid-Hempel there was a 100% infection of colonies investigated¹² (by Crithidia bombi) showing the degree of seasonal variation.
So far, evidence has suggested that bumblebees have been fairly successful in the urban environment, even to the point where some colonies appear to be surviving and foraging over winter in the UK⁴. A study by Osborne et al. suggested that bees preferentially nested in garden locations⁸ and so because urban landscapes (cities such as London, belfast and Cardiff in the UK) are composed of around 25% garden on average,⁹ this might arguably provide a better environment for the bees to thrive in.
However, a study by Banaszak-Cibicka and Żmihorski suggested that not all bees will thrive in the urban environment, with smaller bees being ill suited to the urban environment¹⁴. It may be the case that certain species suffer from increasing urbanisation due to changes in availability of nesting sites and food plants. Therefore, research into the specific interaction between individual species and their environment needs to be conducted.
Part of the decline of bumblebees in recent years has been attributed to the increase in use of commercial and/or foreign bees spreading pathogens to local wild bees both in the UK¹¹ and in the South American continent ¹³.
Parasites themselves have been associated with a number of negative impacts. Crithidia bombi have been suggested to lower bees capabilities for learning flowers and thereby negatively impact foraging habits ¹³ and it has been suggested parasite infection delays worker reproduction¹ which interestingly contradicts the results of this project.
Bumblebee parasites
A parasite is an organism that may live external or internal to the host and obtains all its required nutrition from the host organism, usually in a manner that holds no benefit to the host.
Not all bee parasites have been included in the descriptions below as not all were found in the samples and so are less relevant for this study as their effect cannot be quantified with the data collected in this study. For example, Nosema bombi was not found in any of the bees analysed.
Crithidia bombi
Crithidia bombi are among the most common bumblebee parasites and infect the intestinal tract, specifically the hindgut. They are protozoans and have a flagellum which they use for movement. In fact, in most of the samples some Crithidia could still be seen clearly moving despite being frozen at minus 20°C for over a month.
Crithidia may appear in two different forms, in a pinched spherical shape (promastigote) that one might liken to a lemon and as a more elongated, bacillus shape (choanomastigote). These indicate different life stages of the organism; with the latter being the more mature of the two. The flagellum is present at both and under a phase microscope both appear dark, almost black in colour.
Crithidia may be transmitted horizontally; meaning that if a bee visits the same flower as an infected bee it may also become infected⁷ or may be transmitted through bees within the colony; and vertically when an infected reproducing worker or queen reproduces, the offspring may be infected².
On a microscopic level they are often confused with yeast cells which regularly inhabit the hindgut and Malpighian tubules. These, while being a similar colour and sometimes size are not as consistent in shape as Crithidia and lack a flagellum.
These parasites have been associated with reduced worker reproduction¹.

Crithidia in hind gut	Zoomed in image of Crithidia in hind gut

Apicystis bombi
These are protozoans which reside and reproduce in the fat bodies. When present these were especially noticeable as they occurred in huge frequencies. They were also found in the hind gut and Malpighian tubules. These were found with the second highest frequency in the samples investigated in this project. Interestingly these have spread to native bees in Argentina⁷, thereby widening the global impact of these parasites and posing an even greater threat to bumblebee populations.

Apicystis in fat body	Apicystic in fat body zoomed in

Syntretus
These are the larvae of parasitic wasps which grow inside the host bumblebee. They will eventually kill the host once they are mature and burst from the host unlike all the other parasites investigated.

Syntretus in Abdomen	Syntretus exposed individually

These were found in huge numbers within a single bee, ranging from 11-40. The organisms occupied almost the entire body cavity and had a mass that was easily greater than that of the bee’s internal organs.
Methodology
-Defining regions of different land classes
This project used colonies from a total of 38 different sites of varying urban, agricultural and industrial development. One might note that geographical analysis of some sites were not completed and so 39 of the dissected bees were not used in the final data analysis.
The urbanisation of each site was measured by the use of land classification through the online program QGIS. A radius of 750 meters around each site was used as this is the approximate foraging distance of a bumblebee, making it possible to gain an accurate perspective of the kind of environment they were experiencing during foraging.
The first step was to separate different land use areas by using the ‘split features’ and other tools ‘Split features’ is a tool which allows drawing around an object/area with many straight lines to create a polygon.The aim of this was essentially to divide the site into hundred of polygons, each of which represented a land use type. Roads, woodland and fresh water are but a few examples of the land use classes represented by the different polygons. The resulting outline shown in figure x isolates each division between the different terrains.
This was done to a fairly high degree of accuracy as figure x shown below, for example, had over 800 different features (sections of land class or polygons) meaning that error was minimised. However, there is always a degree of inaccuracy due to the fact that certain areas are shadowed (for example by trees) and by slight inaccuracies in creating shape outlines in feature splitting.
Assigning land class

Example of splitting features (left) Example of abbreviation for assigned land use type (right)

Example site after splitting of features (fig 1A)

The next step was assigning land use types to each of the features for each site. This was done via abbreviations (fig 1A) such as R for road (shown in grey in fig 1A). After assigning a land use type, a colour was assigned to each abbreviation. This led to a total of 30 different land classes. While these were later simplified to impervious/ non-impervious the level of complexity used at this stage will be useful in later projects for which these results will be used.
Once this was complete, the colour scheme was applied to the site. There is a very clear difference between rural village (fig 1D), agricultural (fig 1B), suburban (fig 1E) and urban (fig 1C). Figure 1A and 1E show the same site at the different stages of analysis.

Table to show land use classification
Agricultural land use example (Fig 1B)	Urban land use example (Fig 1C)
Rural village land use example (Fig 1D)	Suburban land use example (Fig 1E)

Urbanisation calculation (impervious surface)
Three of the total 26 sites had an estimated value for the proportion of impervious surface due to time constraints on the project. This was done by comparing the site visually to already calculated values of sites with similar levels of urban development. While this is not the most accurate method, on balance it makes up only a small proportion of the data. While many different land use classes were assigned, the proportion of impervious surface is a recognized way of assessing urbanisation¹⁵ and this was an appropriate division based on the project. The detail of land use classes will be used for a wider project.
The calculation itself simply involved calculating the area of the individual polygons created by the splitting of features (see methodology) and then after determining the polygon to be either impervious or non impervious, calculating the proportion of impervious surface in each site. This is represented below in anequation.
Proportion impervious surface = area of impervious surface/total area
Dissection
Through the course of this project 237 worker bees were screened for parasites. This involved dissection of the bee followed by the screening of samples of the hindgut, Malpighian tubules and fat bodies. This was done using a phase microscope in order to see Crithidia and each sample was checked thoroughly by two people. Photographic evidence of any ambiguous slides were taken and shown to an expert in the field. The slides were immediately sealed with clear nail varnish and put in dry ice or frozen to be preserved in case there was a need to check them later on.
Ovary development and the length of the terminal oocyte were also measured using a graticule. A possible source of error here is that throughout the dissection 25% ringer solution (an isotonic solution that mimics an animals body) was used including on the ovary. While this was necessary in order to move the ovary without damaging it and to see the terminal oocyte, it did however affect the perceived size of the ovaries by refracting the light in particular around the edges of the droplet.
The length of the terminal oocyte was used to measure ovarian development because this is a trusted and frequently used method for assessing ovarian development ^{15, 16} . Worker bees can, contrary to popular belief, reproduce. However, they lay eggs which develop into drones (males) and the oocytes are essentially eggs. As a result, size of the terminal oocyte is essentially a measure of how ready the bee is to produce a drone.
However, it was noted that this might not give a complete picture of ovarian development given that some undeveloped ovaries were clearly more developed than others and so ideally there should be a way to distinguish between these. Furthermore, some very large ovaries had many developed oocytes yet the size of the terminal oocyte did not reflect this.
Each individual ovariole was measured and recorded. If the terminal oocyte was not clearly defined it was marked as U (undeveloped) and where the ovary was developed but degraded it was recorded as D.
Results
Overall, data was collected for ovarian development, presence of parasites and the proportion of impervious surface in the area surrounding each hive. There was no trend found between proportion of ovarian development and proportion of impervious surface (fig 2A) or between the size of the terminal oocyte (not including undeveloped ovaries) and the proportion of impervious surface (fig 2B).

Fig 2A	Fig 2B

There was also no observable trend using the average length with undeveloped (U) as 0 (fig 3). This effectively records undeveloped ovaries as being 0mm in length.

Fig 3

Interestingly when average ovarian length was used on a colony-wide scale and correlated with impervious surface, a slight trend was observable.

Fig 4

There were also little difference in averages between infected and non-infected bees for ovarian length without U and proportion impervious surface.
However, there was a clear difference in average proportion of developed ovaries and average ovarian length with U as 0. These two differences support each other because fewer undeveloped ovaries would produce a greater value for both averages (ovarian length with U as 0 and average proportion developed ovaries). When one considers that the difference in average without U was >1%, it becomes clear that infection does not increase or decrease the length of the ovaries, but instead appears to encourage their initial development.

	Non infected	Infected	Difference	% difference (3 sig fig)
Average ovarian length without U	19.6908076	19.58588	-0.1049276	-0.533
Average ovarian length with U as 0	7.4354047	8.709524	+1.274119	+17.1
Average proportion impervious surface	0.297103278	0.294372381	-0.002730897	-0.919%
Average proportion developed ovaries	0.372306	0.460938	+0.088632	+23.8%

A clear observable trend was visible when the proportion of bees with parasites was correlated with the proportion of impervious surface on a colony wide scale. The trend indicates that more urbanised environments favour lower levels of colony infection.

Fig 5

Furthermore, a clear trend was also seen when proportion of parasites was correlated with proportion of developed ovaries within parasitised colonies. This interestingly suggested that the higher the proportion of parasites, the higher the proportion of developed ovaries. There is a very clear outlier to this trend circled below. One might note that no trend was observable on an individual bee level (fig 2A).

Fig 6

Discussion
Sources of error
The disadvantage of using microscopy to identify parasites is that even with two people checking each sample there is still a risk of missing low-level infections. This is especially true in Crithidia where there was variation in whether the organism was moving. Ideally speaking CPR would have been used, however, it would have been far more time consuming. Microscopic techniques are still respected as a method to identify parasites and are often used in spore counts to identify levels of infection¹⁹ although this was not done in this project.
Additionally, in samples of Apicystis it was possible to identify the presence of the parasites even at lower concentrations in the hind gut and Malpighian tubules before looking at the fat bodies. This indicates that lower infection levels would still be identifiable. This is relevant as the concentration of different parasites varied not only from individual to individual, but also from organ to organ. For example, Apicystis was found almost exclusively in the fat bodies of bees.
Furthermore, many samples were somewhat degraded and so this may have caused inaccuracies in the measurement of ovarian length and possibly in the identification of parasites if the parasites themselves had degraded. One might speculate though that the fact that Crithidia were very clearly alive in some samples would indicate that the parasites themselves are unlikely to have degraded.
Discussion
There are two viable trends that have been identified through this project. Firstly, it would appear that the more urbanised environments favour lower infection proportions within colonies (fig 5). This may be due to more hygienic behaviour as urban bees tend to live in larger colonies and so there may be more of a need for urban bees to display hygienic behaviour. Imhoof and Schmid-Hempel suggested that larger colonies were better able to deal with parasite infections in the long term² which might be the reason for the low proportion of infection in more environments.
Furthermore, parasites are transmitted from flower to flower and so it might be the case that urban bees are less likely to re-visit flowers and thereby are less likely to be infected. This might indicate higher numbers of flowers within a more urban environment as this would lead to a lower chance of re-visiting flowers. Equally this could also come about by larger foraging ranges of bees in the urban environments which would produce similar results. One might speculate that as urban environments are varied in flora ⁸ there are fewer tendencies for repetitive foraging habits.
Secondly there is an observable trend between the proportion of developed ovaries and the proportion of parasites within parasitised colonies (fig 6). This indicates that the higher the proportion of bees infected within a colony the more likely bees were to have developed ovaries. This is further supported by the fact that parasitised bees had a 23.8% higher chance of having developed ovaries over non-infected bees and by the fact that when undeveloped ovaries were counted as 0 the average ovarian size for parasitised bees was 17.1% higher than for non-infected bees. This was also shown through a study by Shykoff and Schmid-Hempel which found that more infected bees had developed ovaries than non-infected¹.
This could be due to a host of reasons. Firstly, it is conceivable that infection triggers a panic response in bees similar to that of the death of the queen bee or the queen becoming unviable. This would lead to worker bees developing their ovaries. Ovarian development in workers is generally linked to what is defined as the competition point where the workers will eliminate the unviable queen¹³. It may be that infection triggers a similar response which would affect overall colony survival.
Secondly, because vertical transmission of some parasites (eg Crithidia bombi) is possible it may be that colonies with an infected queen (which would lead to infected brood) are more likely to have developed ovaries.
Thirdly, higher ovarian development might lead to reduced hygiene habits and so thereby increase the chance of transmission between bees.
Finally, it might be that individual bees which have larger ovaries are more likely to forage more and thereby are more likely to be infected. This would cause colonies which had more developed ovaries to have higher proportions of infection. This could be because workers might be more active in terms of foraging subsequent to the death of the queen which would lead to an increased chance of the worker becoming infected. This is however contradicted by van Honk et al. who indicate that workers with greater reproductive capabilities tend not to forage²⁰.
There is also a slight trend towards greater proportions of impervious surfaces leading to lower proportions of ovarian development (fig 4). This may be due to colony size or due to lower rates of parasite infections. It is important to note that these are trends not strict correlations and did not show statistical significance. This is due to the fact that despite the large initial sample size, there was a relatively low rate of infection and so the sample size for infected bees was relatively small. This issue will most likely be resolved in the full study as only a small proportion of the bees collected were dissected and analysed in this project due to time constraints.
Future studies should look at the effect of other land use types such as agricultural land on the outcome of Bumblebee colonies especially in terms of parasites. As the data showed by no means a perfect trend it may be that one or more specific land use types affects the proportion of infection within colonies. If it were possible to identify this it might help conservation efforts by potentially removing or minimising the effect of whatever the land use type might be.
It would be worthwhile also conducting a similar experiment in a different European country as the high proportion of individual gardens in urban landscapes at around 25%⁹ is almost exclusive to the United Kingdom. It would therefore be interesting to see if bees in a similar climate and in urban environments but with lower access to the small gardens produce similar results.
A study which used more bees would be required to confirm the results of this project as due to the relatively low number of infected bees the number of data points was fairly limited. If the trend implied by this data on the relationship between proportion of impervious surface and proportion of infection within a colony, it would support the concept that urban landscapes tend to positively affect bee populations ^{3, 4, 5, 8, 14}.
Conclusion
While it is challenging to draw any firm conclusions from these results it does highlight parasite infestation in urban colonies as an interesting avenue for research. Further studies with larger data sets are required to confirm the trends highlighted in this study. In particular, whether parasitisation is consistently a greater threat to rural colonies as opposed to urban colonies. This, if true, would add to the growing view that urban environments are fairly advantageous for bees.
With the ever increasing threat to bee survival globally and the related implications it is essential to conduct more research to understand how to best direct conservation efforts. The better we understand what factors are damaging bee populations, the more easily and effectively we can help to prevent them.
Acknowledgments
This project was conducted as part of a Nuffield Research placement and the relevant co coordinator was Dr Amelia Frizell-Armitage. The project was conducted at the Royal Holloway University of London and was supervised by Elli Leadbeater. This project is part of the overall PhD project of Elizabeth Samuelson and was conducted with her assistance.
All parasite images were taken by Elizabeth Samuelson but were of the samples used. All original parasite images were corrected through either 25% or 50% sharpening to better show the parasites and some have been edited to increase contrast again to better show the parasites in question. Graphs and screenshots were taken directly from the data and QGIS and are all my own work.
Bibliography
1. Shykoff, Jacqui A., and Paul Schmid-Hempel. 1991. \”Parasites Delay Worker Reproduction In Bumblebees: Consequences For Eusociality\”. Behavioral Ecology 2 (3): 242-248. doi:10.1093/beheco/2.3.242.
2. Imhoof, B., and P. Schmid-Hempel. 1999. \”Colony Success Of The Bumble Bee, Bombus Terrestris, In Relation To Infections By Two Protozoan Parasites, Crithidia Bombi And Nosema Bombi\”. Insectes Sociaux 46 (3): 233-238. doi:10.1007/s000400050139.
3. Hanley, Mick E., Amanda J. Awbi, and Miguel Franco. 2014. \”Going Native? Flower Use By Bumblebees In English Urban Gardens\”. Annals Of Botany 113 (5): 799-806. doi:10.1093/aob/mcu006.
4. Stelzer, Ralph J., Lars Chittka, Marc Carlton, and Thomas C. Ings. 2010. \”Winter Active Bumblebees (Bombus Terrestris) Achieve High Foraging Rates In Urban Britain\”. Plos ONE 5 (3): e9559. doi:10.1371/journal.pone.0009559.
5. Goulson, D., E. Nicholls, C. Botias, and E. L. Rotheray. 2015. \”Bee Declines Driven By Combined Stress From Parasites, Pesticides, And Lack Of Flowers\”. Science 347 (6229): 1255957-1255957. doi:10.1126/science.1255957.
6. Morandin, Lora A., and Claire Kremen. 2013. \”Hedgerow Restoration Promotes Pollinator Populations And Exports Native Bees To Adjacent Fields\”. Ecological Applications 23 (4): 829-839. doi:10.1890/12-1051.1.
7. Durrer, S., and P. Schmid-Hempel. 1994. \”Shared Use Of Flowers Leads To Horizontal Pathogen Transmission\”. Proceedings Of The Royal Society B: Biological Sciences 258 (1353): 299-302. doi:10.1098/rspb.1994.0176.
8. Osborne, Juliet L., Andrew P. Martin, Chris R. Shortall, Alan D. Todd, Dave Goulson, Mairi E. Knight, Roddy J. Hale, and Roy A. Sanderson. 2007. \”Quantifying And Comparing Bumblebee Nest Densities In Gardens And Countryside Habitats\”. Journal Of Applied Ecology 45 (3): 784-792. doi:10.1111/j.1365-2664.2007.01359.x.
9. Loram, Alison, Ken Thompson, Philip H. Warren, and Kevin J. Gaston. 2008. \”Urban Domestic Gardens (XII): The Richness And Composition Of The Flora In Five UK Cities\”. Journal Of Vegetation Science 19 (3): 321-330. doi:10.3170/2008-8-18373.
10. Otterstatter, Michael C., and James D. Thomson. 2008. \”Does Pathogen Spillover From Commercially Reared Bumble Bees Threaten Wild Pollinators?\”. Plos ONE 3 (7): e2771. doi:10.1371/journal.pone.0002771.
11. Graystock, Peter, Kathryn Yates, Ben Darvill, Dave Goulson, and William O.H. Hughes. 2013. \”Emerging Dangers: Deadly Effects Of An Emergent Parasite In A New Pollinator Host\”. Journal Of Invertebrate Pathology 114 (2): 114-119. doi:10.1016/j.jip.2013.06.005.
12. Shykoff, J. A., and P. Schmid-Hempel. 1991. \”Incidence And Effects Of Four Parasites In Natural Populations Of Bumble Bees In Switzerland\”. Apidologie 22 (2): 117-125. doi:10.1051/apido:19910204..
13. Gegear, Robert J., Michael C. Otterstatter, and James D. Thomson. 2005. \”Does Parasitic Infection Impair The Ability Of Bumblebees To Learn Flower-Handling Techniques?\”. Animal Behaviour 70 (1): 209-215. doi:10.1016/j.anbehav.2004.09.025.
14. Banaszak-Cibicka, Weronika, and Michał Żmihorski. 2011. \”Wild Bees Along An Urban Gradient: Winners And Losers\”. Journal Of Insect Conservation 16 (3): 331-343. doi:10.1007/s10841-011-9419-2.
15. Bloch, Guy, and Abraham Hefetz. 1999. \”Regulation Of Reproduction By Dominant Workers In Bumblebee ( Bombus Terrestris ) Queenright Colonies\”. Behavioral Ecology And Sociobiology 45 (2): 125-135. doi:10.1007/s002650050546.
16. Laycock, Ian, Kate M. Lenthall, Andrew T. Barratt, and James E. Cresswell. 2012. \”Effects Of Imidacloprid, A Neonicotinoid Pesticide, On Reproduction In Worker Bumble Bees (Bombus Terrestris)\”. Ecotoxicology 21 (7): 1937-1945. doi:10.1007/s10646-012-0927-y.
17. Goulson, Dave, Olivier Lepais, Stephanie O’Connor, Juliet L. Osborne, Roy A. Sanderson, John Cussans, Louis Goffe, and Ben Darvill. 2010. \”Effects Of Land Use At A Landscape Scale On Bumblebee Nest Density And Survival\”. Journal Of Applied Ecology 47 (6): 1207-1215. doi:10.1111/j.1365-2664.2010.01872.x.
18. Lu, Dengsheng, and Qihao Weng. 2006. \”Use Of Impervious Surface In Urban Land-Use Classification\”. Remote Sensing Of Environment 102 (1-2): 146-160. doi:10.1016/j.rse.2006.02.010.
19. Ravoet, Jorgen, Jafar Maharramov, Ivan Meeus, Lina De Smet, Tom Wenseleers, Guy Smagghe, and Dirk C. de Graaf. 2013. \”Comprehensive Bee Pathogen Screening In Belgium Reveals Crithidia Mellificae As A New Contributory Factor To Winter Mortality\”. Plos ONE 8 (8): e72443. doi:10.1371/journal.pone.0072443.
20. van Honk, C. G. J., P. -F. Röseler, H. H. W. Velthuis, and J. C. Hoogeveen. 1981. \”Factors Influencing The Egg Laying Of Workers In A Captive Bombus Terrestris Colony\”. Behavioral Ecology And Sociobiology 9 (1): 9-14. doi:10.1007/bf00299847.
21. Morse, Chandler C., Alexander D. Huryn, and Christopher Cronan. \”Impervious surface area as a predictor of the effects of urbanization on stream insect communities in Maine, USA.\” Environmental Monitoring and Assessment 89, no. 1 (2003): 95-127.