Synopsis : Getting good answers involves asking good questions. But remember that the bees have no concept of what is ‘best’, or of the calendar.

Introduction

If you’re just starting beekeeping it’s likely you will have a never-ending list of questions about the somewhat arcane and often perplexing hobby you are embarking on.

And if you’ve been beekeeping for years (or even decades) you might have the same number of questions, albeit somewhat more specialised or esoteric ¹. You’ll also probably be involved in answering some of the questions from less experienced beekeepers.

Getting good quality and appropriate answers broadly depends upon three things:

• who (or what) the question is directed at,
• the wording of the question, and
• whether the answer is simply factual or involves a subjective assessment

Let’s take a simple example …

Q. What is the scientific name of a bee?
A. The scientific name for the order of bees is Hymenoptera, but this encompasses many different species of bees, not just a single type of bee. To specify a particular species, a scientific name would be necessary, for example, the scientific name for a honey bee is Apis Mellifera.

… so, let’s be more specific …

Q. What is the scientific name of a honey bee?
A. The scientific name for a honey bee is Apis Mellifera.

By asking a more specific question you have received a better answer.

Unfortunately, it’s still not completely correct.

There is a convention on the capitalisation of scientific names; the name of the genus (in this case Apis) is always capitalised, but the name of the particular species (mellifera; meaning honey-bearing) should start with a lowercase letter.

The correct answer should therefore be Apis mellifera.

Or what?

You may have noticed the who (or what) in the list above. The ‘or what’ was deliberate because, increasingly these types of questions will be answered, not by humans or beekeepers ², but by a machine.

In fact, the two questions and answers shown above were answered by a machine … in this case one called ChatGPT. In this post, ChatGPT responses are shown in red text. The only reason you may not have heard of ChatGPT ³, released just 2 months ago, was because you’ve been living as a hermit in the Tora Bora cave complex for the last few years.

If ChatGPT is new to you then you’re going to have ample opportunity to experience it – or its successors – in the next few months and years. ChatGPT is described as a chatbot which has been trained through supervised and reinforcement learning. It can do a range of things that are a lot more impressive than not quite getting the scientific name of honey bees correct … it can compose music, write computer code and pass written exams for medicine. It can produce essays and it is causing considerable concern in education, where students are already using it for homework ⁴.

Perhaps I’ll save myself some late night writing by asking ChatGPT to do my homework …

Write a 500 word article for beekeepers about hive tools and Taranov swarm control, remembering to include a handful of poor quality puns and some truly shocking, punctuation.

I asked it to do this … it generates the answer in less than a minute. I’ve added the answer as a separate page as I don’t want to get even more distracted from the topic of the post today.

Factual inaccuracies

ChatGPT does OK on the poor puns (beeeing, buzz-worthy etc.) but less well on the shoddy punctuation, unless you count the Oxford comma.

It does much less well on Taranov swarm control. Indeed, I had started by asking it to “Describe how to conduct swarm control using the Taranov method” and received the following answer:

The Taranov method is a technique used for swarm control in beekeeping that involves manipulating the hive to prevent swarming. The process begins by identifying a colony that is likely to swarm and removing any queen cells, which are the cells that contain developing queen bees. Next, the beekeeper will divide the colony by removing a section of comb with eggs and young larvae, and placing it in a new hive box. This creates a population imbalance in the original hive, which triggers the bees to start raising new queen bees. Finally, the beekeeper will monitor the hive to ensure that no new queen cells are forming, and will continue to remove them as they appear. This method aims to reduce the likelihood of swarming, by keeping the colony balanced and preventing the development of new queen bees.

Which is almost completely wrong. I’m not going to describe Taranov swarm control, but it’s not that. If you want to know (and it is an intriguing method, though not one I have used) I suggest you ask an experienced beekeeper, not a machine. It requires a simple, correct, factual answer.

Subjective or qualified answers

The other thing ChatGPT is not good at is answering questions that require some experience and/or may be subjective. For example …

Q. What is the best type of beehive for a beginner?
A. For a beginner, the Langstroth beehive is considered the best option due to its standardized dimensions, ease of management, and wide availability of equipment and accessories. Additionally, it is commonly used by many beekeepers and has a strong community of support and resources available.

… OK, what about something even more important?

Q. What is the best type of hive tool for beekeeping?
A. A J-shaped hive tool is widely considered as one of the best for beekeeping. It is versatile, durable and can be used for a variety of tasks such as prying apart beehive boxes, scraping propolis and wax, and removing frames.

The first answer actually isn’t bad, particularly if you are in the US. However, the justification is a little odd and could equally well be applied to a National, Dadant ⁵ or Layens hive.

The ‘best hive’-type question is one commonly asked by beginners and one where a better question improves the answer. If you instead ask ChatGPT for the best beehive for a beginner in the UK it recommends a National hive (justified by compatibility).

However, the answer ChatGPT gives to the ‘best hive tool’ question is clearly wrong.

Anyone with even passing experience of opening a hive knows that the claw-type hive tool is by far and away the most practical, comfortable and good looking 😉 .

The ’widely considered’ in ChatGPT’s answer is the giveaway. It’s (obviously) never used a hive tool and so cannot speak from experience.

Flame wars

The examples above are trivial but they do show both the abilities and shortcomings of ChatGPT. But it will get better – more accurate, more factually correct, better at providing qualified subjective answers (and hiding the fact that it has no direct experience of any of the things it is comparing).

You’d better get used to it as it will revolutionise our interactions – direct or indirect – with computers; websites, discussion forums, computer programming, teaching, student assessment and – increasingly – creative work as well.

It’s going to put a lot of people out of work 🙁 .

I’m already out of work, so I don’t feel too threatened, but perhaps it’s also going to eventually replace the beekeeping blogger.

As an aside, I thought a ChatGPT-powered ‘user’ on a beekeeping discussion forum like Beesource in the US, or the BeekeepingForum here in the UK would be – at least briefly – entertaining. Some of the discussion threads on these can get really out of hand, even with strict moderation (a thankless task).

ChatGPT can already interact conversationally, its command of the subject and of English (and, of course, a range of other languages) is already better than many readers/contributors, and it would be a whole lot more persistent in an online argument (though it is currently not particularly creative when it comes to insults).

It’s going to render many of these discussion forums worthless and is already banned from some of the computing forums. The scientific journal Nature has established ground rules for its use – it cannot ‘share’ authorship (!), its use in data analysis must be documented etc.

I already find many discussion forums unrewarding … they might get a lot worse, at least in the short term. Since many beginners use them a lot, I thought it was worth mentioning.

Better answers

Although I’ve been wandering some way off topic there are some important points embedded in the first half of this post.

The question you ask influences the answer you get.

In beekeeping, your local environment and your latitude are particularly influential in the lives of your bees.

I gave a talk this evening on queen rearing. One of the questions was ‘How early in the season can I start?’

That’s a perfectly good and valid question, but answering it requires knowing something about the local climate and colony development.

It also varies from year to year … for example, a cold spring delays things.

Some might simply answer ‘mid-May’ or ‘late-April’ … indeed, questions like that may be asked by someone wanting a calendar-based answer.

Unfortunately, beekeeping isn’t that simple. Beekeepers on the UK south coast can often start queen rearing two months before I can here in north west Scotland.

My answer involved something about drone availability. A drone takes 24 days to develop and a few days after that to become sexually mature. Studies have shown that the peak of drone brood production occurs about one month before swarming (Page and Erickson, 1988), though production starts earlier.

And the correct answer is …

So a better answer is to keep an eye on your colonies, observe drone brood production increasing and – 3-4 weeks later (or perhaps a little before ⁶ ) – start your queen rearing with every expectation the bees will have got the timing about right.

In a cold spring they’ll start producing drones later, at a more southerly latitude they’ll produce drones much earlier than they will in northern Scotland.

Yes, the answer is more difficult to understand than ‘the 19^th of April’, but it’s much more likely to be correct because it is based on an understanding of the biology of the bees.

It’s also likely to be correct most years. You can test this by keeping notes. You’ll then have something to refer back to next year and the one after that, and you’ll be able to answer, with compelling authority, anyone who asks you the same question 😉 .

In about 30 years you’ll be able to review your notes – of drone production and queen rearing successes – and see whether the timing needs revising 🙂 .

In doing that you’ll have completed the transition from seeking a calendar-based response, to understanding the drivers that determine colony development and reproduction, and end up with an answer that is generically applicable, qualified and based upon personal experience.

Expect biased answers

My preference for a particular type of hive tool is based upon personal experience (and unrelated to the fact that I bought 20 of them very cheaply a decade ago). Do not underestimate the importance of personal experience in answering beekeeping questions … or its ability to generate biased, unqualified or even completely incorrect answers.

Ask three experienced beekeepers a question and you’ll get five answers … one will be completely wrong, another will involve ‘brood and a half’ (also wrong … obviously), a third will answer a different question altogether and the final two will express diametrically opposing views about whether the J-shaped or claw-shaped hive tool is ‘best’.

The answer you get is based upon the experience of whoever you ask … and how willing they are to answer.

Be warned, it’s not unusual for the most (usefully) experienced and the most vociferous beekeepers to be different individuals. In fact, it’s not unusual for the most vociferous to be much less experienced than they sound.

For a subject as practical as beekeeping, practical experience is far, far more valuable than ‘knowledge’ gleaned from the internet (after all, you might have been reading something written by ChatGPT).

I know the difference between the Miller and Hopkins methods for queen rearing. I’ve not used either (yet) so I don’t know which is better – either outright, or in particular circumstances.

This can all be a bit overwhelming as a beginner … use your judgement, listen, check some of the answers in a good book ⁷ or a reputable online source, ask a follow-up question.

Nobody knows all the answers and it sometimes feels as though the more knowledge you acquire, the more questions appear.

Answer your own questions; observation and understanding

There is no ‘best’ hive, or ⁸ hive tool. The bees don’t care and – through experience – you’ll find what suits your beekeeping.

It’s likely that the ‘best’ anything in beekeeping – bee, hive tool, hive, smoker, forage, honey, hive stand, extractor, queen excluder etc. – is a meaningless concept.

It’s an irrelevant question as far as bees are concerned. There may be good ones and bad ones, but it’s surprising how tolerant and accommodating the bees – and a beekeeper – can be.

I’d strongly recommend that anyone starting beekeeping ignores articles with the word ‘best’ in them – except perhaps this one.

Ask meaningful questions and look for insightful answers.

The ‘when to start queen rearing?’ is a good – albeit incomplete – question and I suggested how I would (or did) answer it above.

In that example it is really by observing and understanding the bees that you answer the question. I think those are two of the most important skills to acquire as a beekeeper; doing so will always help you get better answers, not least because they help answer them yourself.

Asking an experienced beekeeper gets you part way there but it doesn’t come close to working something out yourself.

Do an experiment

I used the word ‘arcane’ in one of the opening sentences. It means mysterious, obscure or little understood. Despite sounding a little like the word archaic – meaning old fashioned or belonging to an earlier period – it has a totally different etymology. Arcane is derived from the Latin arcānus meaning ‘closed or shut up’, whereas archaic is from the Greek ἀρχαϊκός for ‘ancient’.

Nevertheless, bits of beekeeping are both arcane and archaic.

Sometimes they’re ‘old fashioned’ because experience has shown that a particular method works reliably well, so is promoted and becomes widely used. However, sometimes it’s because ”it’s always been done like that” and everyone unquestioningly follows the approach without asking whether there are other – perhaps better – ways of achieving something.

Let’s take a trivial example … starter strips in foundationless frames.

A foundationless frame is a frame containing no foundation (helpfully, the clue is in the name) . Every foundationless frame you use saves you about £1.40 based upon the current price of foundation. What’s more, because commercial foundation contains miticide residues, every frame you use reduces traces of miticides in your hive.

And, if you read online about making foundationless frames you’ll find lots of descriptions, many of which include instructions to provide a wax, or waxed, starter strip attached under the top bar for the bees to start drawing comb from.

When I started using foundationless frames I – unquestioningly – followed these instructions, cutting 2 cm strips of commercial foundation and nailing them in place in my frames.

And they often fell out … which prompted me to ask a question I should have done in the first place, and to do an experiment.

Ask the bees

Where do bees naturally start drawing comb?

Wherever the swarm ends up. If it’s in a previously unoccupied tree hole, or loft space, there are no convenient strips of wax foundation to be used as guides. The cavity may be completely empty.

Clearly they don’t need a starter strip to work from.

However, the beekeeper does not want the bees to build comb totally haphazardly. We want them to stick within the confines of the frame or it becomes impossible to manipulate. It’s therefore useful to provide them with ‘guides’ on where to start … they might not follow them, but they almost always do.

So, by understanding what the bees do naturally, qualified by our own selfish interests in managing the colony, we can ask the bees what suits them … a strip of wax foundation, a waxed wooden guide or a lollipop stick.

And, having worked out which the bees prefer, we can make a decision based upon what also suits us.

I therefore did a simple experiment. I built a dozen or so foundationless frames, each containing three randomly positioned starter strips under the top bar – one third wax foundation, one third waxed wood and one third plain wooden lollipop stick (or ‘tongue depressors’).

Over the course of a season I used these in a variety of colonies and observed which of the starter strips was preferred … indicated by which the bees chose first, or which they avoided.

And the results are in …

It made no difference at all. There was no correlation between the type of starter strip and use by the bees.

They just don’t care.

So I only now use plain wooden starter strips. They are quicker and easier to prepare, impervious to the steam wax extractor and much more robust. I use them in full hives and in my mini-nucs for queen mating, simply gluing the wood in place.

A little understanding of what the bees do, coupled with a little experimentation and some observation, allowed me work out the answer to the question ’What is the best material for starter strips’ ⁹.

I should note that many other have reached exactly the same conclusion independently. I’m not claiming to have discovered this first, but I did formally test what the bees preferred.

Some ChatGPT creativity to end with

To finish I thought I’d ask ChatGPT to do something creative.

Rather than eliciting incorrect answers about Taranov swarm control I asked it for a poem, in the form of a haiku ¹⁰, about honey bee swarms;

Golden bees swarm forth
Honeyed scent on the breeze
Nature’s sweet bounty.

Not bad 🙂

---

References

Page, R.E., and E. H. Erickson, Jr. (1988) Reproduction by Worker Honey Bees (Apis mellifera L.). Behavioral Ecology and Sociobiology 23: 117–126 https://www.jstor.org/stable/4600197. Accessed February 3, 2023. Note: this might not be accessible online, so you could try this instead – Page, R.E. Jr (1982) The seasonal occurrence of honey bee swarms in north-central California. American Bee Journal 121:266-272.