Tag Archives: health

The Autumn of the Matriarch

I’ve previously commented that weak colonies that build up very slowly in Spring are more trouble than they’re worth. The resources they need – syrup, frames of emerging brood, more TLC – are rarely reflected in the subsequent honey yield.

Quite the contrary, they’re often a lost cause and it could be argued that, from a purely efficiency point of view, it would be better if the colony succumbed during the winter than staggered on into the Spring.

Better still, assuming they’re disease free, use the bees in the autumn by sacrificing the queen and uniting the colony with a strong colony. You’ll boost the latter and strong colonies both overwinter better and build up better the following year.

Do as I say, don’t do as I do.

All the above makes perfect sense, but a combination of sentimentality and ill-placed optimism means that it’s not unusual – in late Spring – to find myself being reminded that “weak colonies that build up very slowly in Spring are more trouble than they’re worth”.

And it’s happened again.

One of my colonies was undersized in late autumn and had built up very slowly this Spring. The queen was a little older than most in the apiary but she’d done well in the past and I thought she might have another season in her. Varroa drops in late autumn and mid-winter had been very low and the bees were beautifully tempered, calm, steady on the comb and a pleasure to work with.

But in the first inspection of the year (10th of May) there just weren’t enough of them. The queen was laying, pollen was coming in, there were no signs of disease and the colony behaviour remained exemplary.

Lagging behind

Comparison between colonies is very informative. That’s why it’s easier to maintain two colonies than one. Other colonies in the same apiary were building up well. By late May I was starting swarm prevention measures on these, using pre-emptive vertical splits.

The small colony was largely forgotten or ignored. I peeked through the perspex crownboard a couple of times and could see they were building up.

Slowly.

I got distracted harvesting the early season honey from other colonies, running out of frames and with more swarm prevention and control. I finally completed a full inspection of the colony on the 17th of June, shortly before the summer solstice and the first official day of summer (so still technically Spring).

Queen failure … not epic, but failure nevertheless

The colony had only a couple of frames of brood and covered a frame or two more than that. The temper and behaviour was still very good. The queen was present and laying. She was being attended by a retinue of workers and not being ignored or harassed.

Failing queen ...

Failing queen …

But she was clearly losing her faculties. Many of the cells contained two or more eggs.

Multiple eggs in cells are often seen with laying workers and sometimes seen when a newly mated queen first starts laying. With laying workers the eggs are often placed on the sidewalls of cells and, as they’re unmated, they develop into drones. The brood pattern is scattered randomly around the frame. With newly mated queens the eggs are usually correctly placed in the base of the cell.

Occam’s razor

The colony was clearly doomed. They showed no sign of trying to replace the queen, without which the future was bleak. I needed to rescue something from the situation. The choice depended on my interpretation of what had gone wrong. The options were:

  1. Queen failure, plain and simple
  2. Laying workers in a colony with a failed queen still present (an unusual situation)
  3. A new, recently mated, queen was also present with the old queen (supercedure)

A thorough inspection of the colony failed to find another queen or any evidence of a recently vacated queen cell. Frankly this didn’t take long, the colony was simply too small to ‘hide’ either of these. Option 3 could therefore be discounted. The presence of another queen would be really important if I was considering requeening the colony or uniting it with a queenright hive – both these are likely to go badly if there was a queen still present.

There was no drone brood at all in the colony and the laying pattern was clustered as would be expected from eggs laid by a queen. Option 2 could therefore almost certainly be discounted. Fortunately again as it’s difficult to requeen a colony containing laying workers. As another aside, I can’t remember seeing a colony with laying workers that also contained a (failed) queen.

That left the most likely explanation for the multiple eggs (and the undersized colony) was the simple failure of the queen. For whatever reason, she was laying at a much lower rate than usual and had started laying multiple eggs in cells. Of the three possibilities, this is the most straightforward. Occam’s razor (William of Ockham, ~1287-1347) is the problem-solving principle that states that the simplest explanation is probably the correct one.

Better late than never

The queen was removed from the colony and it was united over newspaper on top of a strong hive in the same apiary. Two days later the Varroa board underneath the colony was covered in shredded paper indicating that the colonies were united successfully.

Successful uniting ...

Successful uniting …

Which is what I should have done in mid-autumn last year.

Better late than never  😉

A few days later I rearranged the colony, placing the two frames of brood into the bottom brood box and putting a clearer board underneath the top brood box. The resulting single colony, now a bit stronger, will be well-placed for the summer nectar flow and the nine frames of drawn comb vacated by the colony will be reused making up nucs for overwintering.


† Interestingly, I’ve never seen several larvae developing in cells after the multiple eggs hatch. Either the excess eggs or larvae must be removed by workers. I presume this means that the workers can’t count eggs, but may be able to count larvae – not literally of course, but by the amount of pheromones produced presumably. If they could count eggs they’d remove the excess and leave only one, making the identification of laying workers (or a recently mated misfiring queen) much more difficult. Something to be thankful for perhaps? They can, of course, identify the origin of eggs – this process is the basis of worker policing which was touched on in discussion of Apis mellifera capensis, and is of relevance to those using grafting for queen rearing.

Colophon

The title of this post is a corruption of The Autumn of the Patriarch, a book by the Nobel laureate Gabriel García Márquez, written in 1975. The book is about the God-like power and status of a dictator, the General, and the awe in which he is held by the people. Of course, this isn’t the situation in matriarchal honey bee colonies, the structure of which is determined as much – if not more – by the workers, the brood and the circulating pheromones.

Take one for the team

You know it makes sense

You know it makes sense

… would have been a much better title for an interesting recent paper on the impact of Varroa on honey bee colonies. More specifically, the snappily titled “Social apoptosis in honey bee super organisms” (Page et al., 2016 Scientific Reports 6: 27210 doi:10.1038/srep27210) attempts to answer how and why the natural host of Varroa, the Eastern honey bee (Apis cerana), copes with mite infestation whereas ‘our’ bees (Apis mellifera), the Western honey bee, succumbs within 2-3 years without mite-control? The paper is Open Access so you don’t need to pay to read it and you can find it here.

Only the good damaged die young 

The authors demonstrate that A. cerana mite-associated pupae die before they emerge, whereas those of A. mellifera do not. As a consequence of this the mite levels are unable to build up to damaging levels in the colony. Essentially the pupae on which the mites feed die very quickly, meaning the mite also dies. They determined this by uncapping and examining age-matched pupae one day before natural emergence (see below) in Varroa-infested or uninfested colonies. Varroa-associated pupae (upper row in the image below) had all died during pupation.

Infested (above) and control (below) A. cerana pupae

Infested (above) and control (below) A. cerana pupae

In an extension to this study the authors showed that puncturing pupae with a sterile glass needle and then re-sealing the cell (you can do this with gelatin) also results in the pupae dying. The needle used had the same diameter as the chelicerae of the Varroa mite, so this treatment recapitulated the physical damage caused by the mite. Since the needle was sterile it was unlikely that the pupae were dying from exposure to the viruses (or other pathogens) transmitted by the Varroa mite. Instead, it seems that the Eastern honey bee has evolved mechanisms of “self-sacrifice” in response to wounding that result in the death of damaged pupae before the infesting mite has had a chance to multiply. Clever.

Social apoptosis

Apoptosis is the term used by cell biologists to describe a series of events that are also called programmed cell death seen, for example, in virus-infected cells. If a cell detects that it is virus infected, a cascade of signalling events result in it undergoing apoptosis (it dies), so preventing the infecting virus from replicating properly and spreading to neighbouring cells in the organism. Social apoptosis is a similar process, the death of an infected – or infested – member of the superorganism, the honey bee colony.

Immunity is a term meaning ‘having resistance to’, for example immunity to measles due to prior vaccination or infection. Generally, immunity is a reflection of strength of the recipient or exposed to the ‘abuse’ caused by the infectious agent. In contrast, the mechanism described for A. cerana is the opposite of this, instead being a form of immunity through weakness or susceptibility.

A. cerana has additional resistance mechanisms that help it combat Varroa infestation including enhanced grooming, removal of mites from unsealed brood, entombing multiply mite-associated drone brood (it’s not clear to me whether this is the same mechanism as the social apoptosis reported here), increased hygienic behaviour and shorter developmental cycles. These will have evolved over the millennia that the mite and bee have associated.

Any chance A. mellifera will evolve a similar mechanism?

Possibly, but I’m not holding my breath. There are already hygienic strains of A. mellifera – for example, VSH bees developed by the USDA group at Baton Rouge. These typically uncap and discard Varroa-associated pupae. This isn’t the same process as the social apoptosis reported here in A. cerana. The latter pupae die prematurely, thereby preventing mite reproduction. While we’re on the subject of Varroa and genetic resistance – do VSH A. mellifera strains open and discard mite-associated pupae … a) early enough to prevent significant levels of mite replication, and b) without releasing progeny mites from the cells they were raised in? I’m aware of the rates at which they clear out Varroa infested cells, but not either the timing of these events or the fate of any Varroa released at the same time.

It’s difficult to imagine a practical strategy to select for A. mellifera honey bee pupae that are more sensitive to Varroa infestation … our bees are currently too robust.


Billy Joel wrote Only the good die young which appeared on his 1977 album The Stranger. “[Not] so much anti-Catholic as pro-lust” Joel explained when it was censored, inevitably ensuring its chart success. The song has more to do with the birds and the bees …  😉

 

Keep your distance

A recent paper by Nolan and Delaplane (Apidologie 10.1007/s13592-016-0443-9) provides further evidence that drifting/robbing between colonies is an important contributor to Varroa transmission. In the study they established multiple pairs of essentially Varroa-free colonies 0, 10 or 100 metres apart and then spiked one of the pair with a known number of Varroa. They then monitored mite build-up in the paired colonies over several months. By comparison of the relative mite increases in colonies separated by different distances they showed that the more closely spaced, the more likely they were to acquire more Varroa, presumably through robbing or drifting.

This isn’t rocket science. However, it’s a nicely-conducted study and emphasises the importance of colony spacing on the transmission of phoretic mites between infested and uninfested colonies – through the normal colony activities such as robbing and drifting – as a primary cause of deformed wing virus (DWV) disease spread in the honey bee population. The paper only studies mite levels, but the association with DWV transmission is well established and unequivocal.

Related studies on the influence of colony/apiary separation

The introduction to the paper provides a good overview of the prior literature on the impact of drifting on disease and Varroa transmission, some of which has already been discussed here. However, some of these studies have not previously been mentioned and deserve an airing, for example:

  • Sakofski et al., (1990) showed that there was no difference in mite migration between colonies in closely-spaced rows from those located up to 10m apart.
  • Frey and Rosenkranz (2014) showed that high-density colonies (>300 within flight range [2.5 km] of the sentinel colonies) experienced approaching 4-fold greater inbound mite migration than when located in areas containing a low-density of treated colonies. Over a 3.5 month period the difference was 462 +/- 74 vs. 126 +/- 16 mites. This would have a very significant impact if allowed to subsequently replicate in the recipient colonies.
  • Frey et al., (2011) previously investigated mite transfer between colonies located 1m to 1500m apart. Strikingly, in this study (which was conducted during a dearth of nectar) mite transmission was effectively distance-independent, with the recipient colonies acquiring 85 – 444 mites over a 2 month period.
Frey and Rosenkranz (2014) Mite invasion ...

Frey and Rosenkranz (2014) Mite invasion …

What can we conclude from these studies?

  1. Closely-spaced colonies – for example, the sort of distances used to separate colonies in an apiary – should really be viewed as a single location as far as mite infestation is concerned. A single heavily-infested colony in an apiary will quickly act as a source of mites to all other colonies.
  2. High densities of beekeepers – assuming the usual range in both the timing and vigour with which Varroa control is practised – is probably detrimental to maintaining low mite levels in your own bees.
  3. Significant mite transmission occurs over distances of at least 1.5 km … not just between hives in a single apiary. How many colonies are there within 1.5 km of your own apiary? Even if you are careful about controlling mite levels, what about all the beekeepers around you?
  4. Colonies wth uncontrolled levels of mite infestation, abandoned colonies (or swarms that occupy abandoned hives) and feral colonies located at least 1.5 km away are potential sources from which your carefully-maintained hives get re-infested …

Recent experience with high and low density beekeeping

One mile radius ...

One mile radius …

I’ve moved in the last year from the Midlands to Fife. Beebase and my involvement with local beekeepers suggest that these represent areas of high and low colony-density respectively. For comparison, Beebase indicates that there were over 230 apiaries within 10 km of my home apiary in the Midlands and that there are currently 20 within a similar range in Fife. In the Midlands I was aware of at least 25 colonies (in several different apiaries) within a mile of one of my apiaries. Furthermore, apiaries might contain lots of hives … one of those previously within 10 km of my home apiary was our association apiary which held up to 30 colonies from ~15 beekeepers. In contrast, the closest beekeeper to my current home apiary is almost 3km away … though I acknowledge there may well be hives “under the radar” belonging to beekeepers that are not members of the local association or have not bothered to registered on Beebase (why not?). It’s far too early to be definitive but mite levels in my colonies have been reassuringly low this season. This includes uncapping hundreds of drone pupae – the preferred site for Varroa to replicate – without detecting a single mite. I’d like to think this was due to timely and effective Varroa control, but it is undoubtedly helped because my neighbours are further away … and perhaps better at controlling the mite levels in their own colonies.

This study provides further compelling evidence of the importance of either keeping colonies isolated (which may not be possible) and ensuring that all colonies in the same and adjacent apiaries are coordinately treated during efforts to control mite numbers.

Gaffer tape apiary

Gaffer tape apiary …

Save the bees, save humanity

I’ve used this poster in talks a couple of times to make a distinction between colony collapse disorder (CCD) in the US and colony losses due to disease in the UK.

Save the bees ...

Save the bees …

It’s a rather striking poster … although it carries the website address www.nrdc.com (which appears to belong to the National Realty and Development Corp.), the logo and the subject are much more likely to be associated with the Natural Resources Defense Council (www.nrdc.org). Whatever … the message is clear, without bees there will be pollination shortages for many important and valuable fruit and vegetable crops. The term CCD, a still incompletely understood phenomenon where hives are abandoned by workers, was first used in 2006 in the USA and similar types of colony losses have been reported in a number of European countries, though not in the UK. Prior to 2006 there were a range of other names given to apparently similar phenomena – spring dwindle, May disease, fall dwindle disease [PDF] etc.

The ‘Save humanity’ statement possibly refers to the the apocryphal quote attributed to Albert Einstein “If the bee disappears from the surface of the earth, man would have no more than four years to live” … though it’s highly unlikely Einstein ever actually said this. It’s also a rather questionable statement. Certainly honey bees provide important pollination services, but so do many other insects (and not just insects). There are certain crops for which honey bees are important – such as almonds – at least on the scale they grow them in California. However, on a visit-by-visit basis, honey bees can be relatively poor pollinators. For examples, solitary bees such as Osmia sp. are much more efficient pollinators of apples. The inefficiency of honey bees is more than compensated though by their numbers and our ability to move hives to crops that need pollinating.

So, if honey bees are so important, why does the picture above show a wasp?  😉

 

 

When to treat?

Preparing Apiguard

Preparing Apiguard …

When and how do you treat colonies to have the greatest effect in minimising Varroa levels? At the end of this longer than usual post I hope you’ll appreciate that this is a different – and much less important –  question than “When is the best time to treat?”.

You probably use one of the treatments licensed and approved by the Veterinary Medicines Directorate (VMD), which include Apistan, Apivar, Apiguard, MAQS and Api-Bioxal. I’ve discussed the cost-effectiveness of these treatments recently. If used correctly, all exhibit much the same efficacy, reducing phoretic mite levels by 90-95% under optimal conditions. That being the case the choice between them can be made on other criteria … the ease of administration, the cost/treatment, the likelihood of tainting the honey crop, the compatibility with brood rearing, whether they mess up your vaporiser etc. After using Apiguard for several years, with oxalic acid (OA) dribbled in midwinter, my current preference – used throughout the 2015 season – is OA sublimation or vaporisation. This change was based on four things – efficiency, cost, ease of administration and how well it is tolerated by a laying queen. The how? you treat is actually reasonably straightforward.

When, not how, is the question

DWV symptoms

DWV symptoms

OK, but what about when? Because, if the treatments are all much of a muchness if used correctly, the when is actually the more important consideration. When might be partly dictated by the treatment per se. For example, Apiguard needs an active colony to transfer the thymol throughout the hive so the recommendation is to use it when the ambient temperature is at least 15ºC (PDF guidance from Vita). It’s worth stressing that this is the ambient temperature, not the temperature in the colony, which in places will be mid-30’s even when it’s much colder outside. At low ambient temperatures the colony becomes less active, and in due course clusters, meaning that Apiguard is not spread well throughout the colony, and is therefore much less effective. If you’re going to use Apiguard you must not leave treatment too late.

For readers in Scotland it’s interesting to note that the SBA annual survey by Peterson and Gray shows significant numbers still use Apiguard in September and October, months in which the mean daily maximum temperature is ~14°C and 11°C respectively … so the average daily temperature will be well below the recommended temperature for effective Apiguard use.

However, the when should be primarily informed by the  why you’re treating in the first place. It’s not really Varroa that’s the problem for bees, it’s the viruses that the mite transfers between bees when it feeds on developing pupae that cause all the problems. Most important of these is probably Deformed Wing Virus (DWV), but there are a handful of other viruses pathogenic to bees that are also transmitted. DWV causes the symptoms shown in the image above … these bees are doomed and will be ejected from the hive promptly. However, although apparently healthy (asymptomatic) bees have low levels of DWV, it’s been shown by Swiss researchers that DWV reduces the lifespan of worker bees, and that high levels of DWV in a colony are directly associated with – and causative of – overwintering colony losses. Therefore, the purpose of late summer/early autumn treatment is to reduce the Varroa levels sufficiently so that high levels of the virulent strains of DWV are not transmitted to the overwintering bees. When? therefore has to be early enough that this population, critical for overwinter survival, will live through to the spring – however long the winter lasts and however severe it is. However, before discussing when winter bees are reared it’s worth considering what happens if treatment is used early or late.

What happens if you treat early?

Mid June

Mid June treatment …

For example, mid-season or after the first honey crop comes off. Nothing much … other than slaughtering many of the phoretic mites. This is what most beekeepers would call “a result” 😉  Aside from possible undesirable side effects of treatment – like tainting honey, or preventing the queen from laying or even, with some treatments, queen losses – early treatment simply reduces mite levels. It’s important to remember that the levels may well not be reduced sufficiently to negate the need for a treatment later in the season … as long as there is brood being raised the mites will be reproducing (for example, look at the mid-June treatment generated using BEEHAVE modelling – image above). Furthermore, avoiding those undesirable side effects might require some ‘creative’ beekeeping (for example, clearing the supers and moving them to another hive) and will certainly inform the choice of treatment but, fundamentally, if the mite levels are high then treating earlier than is usual will benefit the colony, at least temporarily. If the mite levels – estimated from the disappointingly inaccurate mite drop perhaps – are dangerously high you should treat the colony.

What happens if you treat late in the season?

Isolation starvation ...

Isolation starvation …

In midsummer workers only live for ~40 days. If mite levels are high, virus transmitted to these workers will shorten their lives, so reducing the colonies’ foraging ability and – possibly – ability to defend itself against wasps or robbing late in the season. However, if you delay treatment until very late the lifespan of bees raised at the end of the season – the overwintering bees – will be reduced with potentially more devastating consequences. The usual winter attrition rate of workers will be higher. The cluster size of the colony will shrink faster than a colony with low mite levels. At some point the colony will cross a threshold below which it becomes non-viable. The cluster is too small to move in cold periods to new stores, resulting in the beekeeper finding a pathetic little cluster of bees in a colony that’s succumbed to isolation starvation. A larger cluster, spread across a greater area and more frames, is much more likely to span an area of sealed stores and be able to exploit it.

When are winter bees reared?

Frosty apiary

Frosty apiary

In the Swiss study referred to above they looked at the longevity of winter bees. The title of the paper is “Dead or alive: deformed wing virus and Varroa destructor reduce the life span of winter honeybees”. We can use their data to infer when winter bees start to be reared in the colony and when mite treatments should therefore have been completed to protect these bees. Their studies were conducted in Bern, Switzerland, in 2007/08 where the average temperature in November/December that year was 3ºC. They first observed measurable differences in winter bee longevity (between colonies that subsequently succumbed or survived) in mid-November. This was 50 days after bees emerged and were marked to allow their age to be determined. By the end of November these differences were more pronounced. Therefore, by mid-November Varroa and virus-exposed winter bees are already exhibiting a reduced lifespan. Subtracting 50 days from mid-November means these bees must have emerged in late September. Worker development takes ~21 days, so the eggs must have been laid in the first week of September, and the developing larvae capped in mid-September.

To protect this population of overwintering bees in these colonies, mite treatments would have had to be completed by the middle of September, so that mite levels were sufficiently low that the developing larvae weren’t capped in a cell with a Varroa mite carrying a potentially lethal payload of DWV. For Apiguard treatment (which takes 2 x 14 days) this means treatment should have been started in mid-August. For oxalic acid vaporisation (which empirical tests suggest is best conducted three times at five day intervals) treatment would need to start no later than early September and preferably earlier as it is effective for up to a month.

Of course, these figures and dates aren’t absolute – the weather during the study would have influenced when the larvae would be raised as winter bees, with the increased fat deposits and other characteristics that are needed to support the colony survival through the winter. Despite the study being based in Switzerland my calculations on dates are probably broadly relevant to the UK … for example, the temperature during their study period is only about 1ºC lower than the 100 year average for Nov/Dec in Eastern Scotland where I now live.

In conclusion

That was all a bit protracted but it hopefully explains why it’s important to be selective about when you administer Varroa treatments. Chucking in a couple of trays of Apiguard in mid-August or mid-October has very different outcomes:

  • in mid-August the phoretic mite population should be decimated, reducing the transmission of virulent DWV to the all-important winter bees that are going to get the colony through the winter. This is a good thing.
  • in mid-October the mite population will be reduced (not decimated, as it’s probably too cool to effectively transfer the thymol around the hive – see above) but many of the winter bees will already have emerged, probably with elevated levels of DWV to which they will succumb in December or January. This is a bad thing.

Perhaps perversely, treating early enough to prevent the expected Varroa-mediated damage to developing winter bees is not be the best way to minimise mite numbers in the colony going into the winter. Using BEEHAVE I modelled the consequences of treating in the middle of each month between August and November¹. I used the default BEEHAVE setup as described previously. Figures plotted are the average of 3 simulations, each ‘primed’ with 20 mites at the start of the year.

Time of treatment and mite numbers

Time of treatment and mite numbers

There’s a lot on this graph. To show colony development I plotted numbers of eggs, larvae and pupae (left axis) as dotted red, blue and black lines respectively. Mite numbers are shown in solid lines – treated with a generic miticide in mid-July (black), mid-August (blue), mid-September (brown), mid-October (cyan) and mid-November (green). In each case the miticide is considered to be 95% effective at killing phoretic mites. The gold arrowhead indicates the period during which winter bees are developing in the colony, based upon the data from Dainat.

Oxalic acid trickling

Oxalic acid trickling

Treating at or before mid-August controls the late-summer build up of mites in the colony – look how the blue line changes direction. Mites that are not killed go on to reproduce in late September and early October, resulting in levels of ~200 at the year end. Remember that mites present in midwinter can, in the absence of sealed brood, be effectively controlled by trickling or vaporising oxalic acid (Api-Bioxal), and that this Christmas miticide application is particularly important if the autumn treatment has not been fully effective. In contrast, treating as late as October and November (cyan and green lines) exposes the developing winter bees to the highest mite levels that occur in the colony doing the year, and only then decimates the phoretic mite numbers, with those that remain being unable to reproduce effectively as the brood rearing period is almost over. Starting treatment in mid-September isn’t much different, in terms of exposing the winter bees to high mite levels, than starting later in the year.

So, within reason, treating earlier rather than later both reduces the maximum mite levels and helps protect the winter bees from virus exposure. Of course, treating as early as mid/late August may not be compatible with your main honey crop (particularly if you take hives to the heather) … but that’s another issue and one to be addressed in a future post.

STOP PRESS There is a very important follow-up article to this. Kick ’em when they’re down describes why it’s so important to treat during a broodless period in midwinter to minimise mite numbers at the start of the following year. Just treating in late summer is not sufficient … you’ll protect your winter bees, only for them to be targeted by mites the following Spring.


¹BEEHAVE makes a distinction between ‘infected’ and ‘uninfected’ Varroa, the proportions of which can be modified. This might (no pun intended) not accurately reflect the reality in the hive, where Varroa-mediated transmission of DWV results in the preferential amplification of virulent strains of the virus. I need to roll my sleeves up and delve into the code to see if the model can be altered to fully reflect our current understanding of the biology of the virus. This might take quite a while …

References

Overwintering honey bees: biology and management from the Grozinger lab.

Managing Varroa (PDF) by the National Bee Unit