A Night on Robben Island 

Robben Island is known as the place where Nelson Mandela was incarcerated for 18 of the 27 years he spent incarcerated. But Robben Island, now a UNESCO World Heritage Site, has been used as a leper colony, an animal quarantine station and a prison island. During World War II, the island accommodated guns to sink potential invading ships.

Robben Island has amazing wildlife. Although named after a Cape fur seal colony, there are now only a few around. But there are lots of penguins (seemingly everywhere), seagulls and many other marine birds. As well as the amazing indigenous wildlife, Robben Island has a darker invasive side. The population of rabbits that once turned most of the island into a sandy desert have been reduced to a very few individuals (we didn’t see any), cats are nearly gone, but herds of fallow deer are still impacting what remains of the native vegetation.

From Top Left to Bottom Right: View of Table Mountain from the boat, doing another tour of the island but during the day, a WWII lookout installation, Ben behind bars for the night.

CIB visiting fellow Ben Allen and I were allocated a prison cell for the night to observe some of the animal control program in action. Ben has worked on control of mammals in Australia for many years, and he had some interesting ideas for helping to improve the control effort on Robben Island. The native birds and steenbok certainly make the control effort far more complex, but we believe that given appropriate effort to remove the last individuals, lasting control is possible.

Thanks to the Robben Island museum for making the trip possible.

Are you sitting comfortably? Then I'll begin

I started thinking about this topic some years back as I often need to write popular articles that make some of the science that we do more generally accessible. But then a few days ago, an article popped up in The Guardian (see here) by Nick Enfield that made me think again.

“No story telling” is a comment that I sometimes make when reading drafts of manuscripts, chapters, and even when editing for journals. What do I mean by this? Stories are deterministic. That is to say that the story teller has an end in mind when they start telling the story, and the telling is a way to get to their goal. A story that’s ‘pointless’ will frustrate the audience and won’t engender them to listen to that storyteller again. In a good story, reaching that goal will often result in lots of twists and turns with the goal shrouded in mystery until it is revealed. In a teaching story (like a parable), the goal may be overt, such that the audience relates to the narrative and buys in to the same conclusion.

If we did science like we tell stories, we would decide on the way the system works before we studied it, and then design the experiment in order to reach our desired goal. You should have recognised by now that this is not the way we do science. This is clearly an undesirable way to go about doing science because we should never prejudice the result that we’ll get from a study before we do it.

We need to approach science in a very different way to storytelling. When I was doing my PhD I was very frustrated as I had the impression that my supervisor knew what result he wanted and designed the study to show it. This is known as “confirmation bias”. When an experiment failed to meet the expected result, he declared that it had failed.

In our studies, we read other studies and observations to formulate a question that we frame as a hypothesis. We then devise an experiment that will test this hypothesis in the most objective way, so that we can fairly accept or reject our null hypothesis (see blog on hypothesis forming here).

Thus our studies are the very opposite of story telling. Or are they?

I often tend to think of the answer to the question as the goal of the study. That I don’t know what the answer is, doesn’t spoil the story for me. The important thing is asking the question. An unexpected answer might send us back to thinking more about the system that we are studying and result in a greater revelation.

A really good example of this is the study published earlier this year by Becker et al (2018). Francois found a strong relationship between probability that Rose’s dwarf toadlets survive and rainfall in the breeding season. More rainfall equals greater survival is what one would automatically think, but that wasn’t the result obtained. Francois found that survival increased with less rainfall. It wasn’t until we made the connection between the fact that increased rainfall during the breeding season meant that the toadlets spent more time in puddles, decreasing their survival. As the toadlets aren’t feeding during this time they lose weight, and also expose themselves to more predation pressure. In a dry year, the toads will head back to their subterranean refuges much earlier and continue to pursue ‘safer’ feeding and hiding habits. While we might intuitively feel that a ‘safe toadlet’ is a better life-strategy, reduced rainfall means reduced reproduction, and so results in a failure for toads that don’t manage to pass on their genes that year. The result is a variable life-history with the weather, something that was previously unknown. To me, that’s a great story!

But the scientific paper that was written about this ‘story’ (Becker et al 2018) doesn’t have a narrative style, and doesn’t fit the description of a narrative that we discussed above. Instead, it follows the formula that we set out way back in the blog (Introduction, Materials & Methods, Results and Discussion: see here). This style does not treat the experiment as a complication on the way to the story’s goal. The structure introduces the rationale behind conducting the experiment, then objectively explains the findings, and lastly discusses their meaning in relation to what is already known. The key differences in the storytelling style from the scientific formula are the absence of a known goal at the start of the process (see a more detailed discussion of this by Yarden Katz here).

But there is a role for stories when communicating science as this increases interest, facilitates understanding, and enhances memory. This is particularly true when communicating science to the public and making it more accessible, but it also applies to interactions between scientists, for example at conferences. The presentations that tell a story and entertain are those we tend to remember. Not easy, but if we do want to communicate well with each other, then we need to learn the art of story telling, without compromising our scientific objectivity.  

Which type of alcohol is the best?

We use alcohol for all sorts of things in biological sciences, but alcohol isn't always what it seems, and there are some important dos and don'ts. I got the idea and some of the facts for this blog post from posting by Cristy Gelling in 2012 that’s been sitting on my computer for 6 years. I’ve passed it on many times, but realised that I should do a blog post on it, because it’s useful information like this that you’re supposed to have on the lab blog!

This is the chemical make-up of ethanol. I hope that it’s familiar to all of you as the two carbon molecules have their free spaces taken up with hydrogen, except one that has the -OH ending making it an alcohol. This is ethanol C₂H₅OH, because there are two carbons. One carbon would be methanol CH₃OH, and more on that later.

Ethanol is familiar to most of us because it’s the same intoxicant that most of us use on ourselves when we’re trying to loosen up socially. It turns out that this really is a very poor choice of social intoxicant, because it is highly addictive and causes all sorts of diseases and problems, including the same social situations that we were hoping it’d make better. If you don’t do alcohol, then good for you, but if you do please make sure that you are very careful in how you poison yourself with this toxin as it’s really very dangerous. And it’s really important to say that you must never use lab ethanol to dose yourself (or anyone else). The reason is that lab ethanol is often close to being pure, and at those levels can do real harm. Moreover, lots of alcohol that’s kept in the lab isn’t ethanol, so you can do a lot of damage even when you were thinking…

Why do we use ethanol so much?

We use ethanol for pickling specimens (preserving bodies after formalin fixing), and for keeping tissue samples so that we can extract whole DNA later for molecular studies. The reason why it’s really useful is that it gets rid of water – dries the sample out. Water is really bad news for DNA, and other tissues. You should remember that water is a polarised molecule (one side is negative and the other positive). This means that it is very destructive and once an animal (or a piece of an animal) is dead, the water in the tissue will start destroying all of the molecules inside that we’re interested in. Very soon the rot will set in.

Ethanol grades:

95.6% ethanol

If your ethanol is locally distilled, then this is what you actually have. You aren’t going to have 100% ethanol through distillation because at this point the distilled ethanol has reached its vapour point (azeotrope) such that the vapour state has the same ethanol:water ratio as the liquid state.

A lot of people use ethanol for cleaning. This is because it’s a really good solvent – better than water at some things, and water is an awesome solvent. Oddly, I often see people cleaning benches with ethanol. This is really just moving the contaminants around on your bench, and not actually killing them. If that’s your objective then fine. Otherwise, use 10% bleach.

Absolute ethanol (99-100%)

This is expensive stuff, because as you’ve read above, it’s not as easy as distillation to get it. To get rid of that last 4.4% of water, the chemists use additives (such as benzene) to purify it by disrupting the azeotrope. You really only need absolute ethanol if you are doing something that has sensitivity to water (i.e. there must be no water), that calls for Analytical Reagent Grade (ANALR). Otherwise, use 95% which is all you’ll need for most of your lab work.

Remember that absolute ethanol is super hygroscopic (attracts water), so if you leave the lid off, you’ll won’t have absolute ethanol for long. For most of our work, best leave it on the shelf.

Preserving alcohol for specimens (70%)

Because water is naturally present in animal tissues, completely drying the specimen out in 95% ethanol will cause shrinkage and the tissues become brittle. Thus we use 70% alcohol when preserving specimens for long periods. Remember, that the preservation is about carefully managing the water content of the specimen. If you place a large fresh toad in 70% alcohol, the alcohol will become diluted by the water in the toad. The volume that you put the specimen in is critical. A large specimen in a small jar is mostly water before you’ve added any alcohol.

The best way to think about a specimen is a balloon full of water (most animals are 98% water). You need a large enough volume of water so that the volume of the balloon doesn’t change the concentration of the alcohol in the jar. Well, of course it will do exactly this unless you have a massive jar! Therefore, after a few days of the first soak, the alcohol must be changed. Otherwise, your specimen will rot. Well curated museum specimens will have their alcohol changed a few times in the first year, and then topped up on a regular basis.

Not everyone wants their specimens in 70% alcohol. This is because different taxa do best in different alcohol pickling states. If in doubt ask. For most amphibians, we want the final concentration to be 70% (but note that this isn’t what it’ll be when you first put the specimen and alcohol together).

How do you make 70% ethanol?

Just add 30% volume of distilled water to 70% volume of 95% ethanol. I know that it’s not exact, but this will be fine (for the dilution issue stated above). Never use absolute ethanol, because it’s super expensive. And it’s super expensive because someone has spent lots of time and effort removing all of the water. So you’d be a real imbecile to tip water into it.

Also remember that over time, 70% ethanol loses the alcohol (through evaporation) but keeps the water. So that really old bottle of 70% that’s been sitting on the shelf for years probably isn’t any more. If in doubt, make up some more.

What about denatured alcohol – methylated spirit – rubbing alcohol?

If you are out in the field and you need alcohol for work, the only thing that you are likely to be offered is denatured alcohol. This is because shops often aren’t allowed to sell 95% alcohol because folk are likely to do stupid things with it, as I’ve already warned you about above. If you want to know more about what’s inside, you can read it on Wikipedia. But is it useful to us?

For preserving bodies, denatured alcohol does a pretty good job. But make sure that you let the museum curator know that that’s what you’ve used as they’ll want to thoroughly rid the specimen of the denatured alcohol before putting it in their ethanol collection. However, some collections use denatured alcohol for all of their specimens. Thus, if you’re borrowing specimens to work on, it’s important to find out what the museum uses in order for you not to make a mistake when you top up the jar. If in doubt, do not use denatured alcohol.

Denatured alcohol is bad news for DNA, as the additives can interfere with the extraction and other applications (e.g. fluorescent labels used in microsats). You will find that some people do use it without problem. It maybe that it’s fine for your purpose, but if you have no idea don’t risk it. Use 95%.

Methanol

It’s not likely that you’ll need to use methanol. Methanol is a poison, so if you have to use it, treat it as such. It’s what people produce when distillation is incomplete, and has caused large numbers of people to go blind (extreme), or get really bad headaches (common). Methanol is often used in the denaturing process.

Why should an editor read your submission?

There is a worrying increase in poor editorial decision making, without any basis, because editors are not reading submissions.

When a manuscript is submitted to a journal, the submission goes to either the editor-in-chief or a handling editor based on the key words or journal section implied during submission. In some journals (like PeerJ) the submissions are offered up to a whole group of editors who can take their pick. It seems that the next thing that happens is that the manuscript is sent out for peer review. But stop. That’s not correct and it’s really not a good way to proceed. Before sending it out, the designated handling editor needs to read the submission.

Why is reading so important?

The title and abstract really don’t allow a handling editor to decide whether or not a manuscript should go out to review. There are a lot of manuscripts out there that should not have been submitted, because their authors do not have sufficient judgement of their own or because they believe that there is a reason to just ‘chance it’. It is very important that handling editors read the submission, because without that they are moving editorial responsibility from themselves to the peer reviewers.

Some years ago, I co-authored a series of articles (see here) that were published across many journals about how peer review was becoming very difficult for editors because so few colleagues would accept to do reviews. This was a problem then, and it’s still a problem now. I’ve recently sent out manuscripts to more than 15 people before getting two reviews. That peers are not prepared to review, or in many cases even to respond to the request, is very poor. However, more recently I’m experiencing a sharp increase in manuscripts to review that should never have been sent out.

My time is precious, and it’s becoming quite expensive for my employers. I am happy to conduct peer review because it is an important part of the scientific publishing process, and I expect others to review my own work. However, I expect that any manuscript that I receive is worthy of my attention and time. If the handling editor has not read it, they cannot decide this and I really wonder what makes them think that they can send it to me (and presumably others) to read while they don’t feel that they have time to do it themselves. Moreover, this appears to be a trend among younger less experienced editors (often associates) that have either not received any guidance in what their job as editor it, or they should not be editors.

If you are going to be an editor, then you must be prepared to read

I must admit that I’ve done it. I’ve sent out manuscripts to be reviewed as I didn’t have the time to properly read the article, but a superficial skim suggested that it seemed fine. Not good. It’s embarrassing to handle manuscripts that should be rejected without peer review. In the case I’m thinking of, once I’d read through the manuscript later on that day, I realised how bad it was and immediately wrote to those I’d asked to do the reviews and asked them not to. The article was rejected. I only do this if there’s no science contained therein. It’s horrifying how often that’s the case, but I’d rather take on this burden as an editor than burden two or three times as many others to make the same call.

Sometimes, it’s not clear whether or not a manuscript will pass muster. Articles can stand or fall on good or bad single judgements of the authors. But misjudgements aren’t always obvious to editors. That’s why peer review is important, and that’s why it’s hugely important for editors to send manuscripts to appropriate reviewers that have some expertise in a subject. For example, if I receive a manuscript about the calls of East Asian frogs, I shouldn’t only send it to people who work on African frogs. It’s really important that someone familiar with the animals reads the manuscript. This is because they might know something that others would miss. If they spot an error in the identification of the species call in the manuscript, the entire premise of the science might fall apart.

As I’ve discussed before (see blog post here), science is built on the work that others have done before, but basing your work on what someone else has written will mean that you have a good understanding of what they have done and how they have done it. Assumptions have to be made to get anything done, and it’s a good exercise to sit down with a published paper (or even a manuscript or a colleague or your own) and read through listing all the assumptions that are made. Physicists might have a very long list if they read a biologists manuscript, but with some practice you learn to see the assumptions that the authors have made when designing their experiment, or going out to the field to conduct their study. An incorrect assumption could lead to the entire manuscript losing its value.

In my example above, the authors might assume that they had correctly identified the species when recording its call. Such assumptions should be backed up with museum and/or tissue bank accessions. But when they are not, the assumption that the authors are recording what they think they are, is vital. If this is placed in doubt, then the entire premise (description of a call to distinguish this species in the field) simply falls apart. In a case without vouchers, the assumption needs backing up by someone who knows the identification from another study, or without any foundation it becomes worthless.

I’ve been on the other end too

I’ve submitted manuscripts to journals where the editor clearly never read the manuscript. Editors who have made a decision without any guidance of their own gives this away. If your decision comes as a single sentence that asks you to revise according to the reviewers’ comments, then you can be reasonably sure that your editor hasn’t read the manuscript (and possibly not even the reviews).

It’s not surprising that the editors have little to nothing to say; without reading the manuscript, the reviewer comments aren’t really very helpful. Without reading, the editor has no idea whether the reviewer is biased or (as is sometimes the case) deluded. As an editor, you simply have to read. And if you don’t have time to read, you shouldn’t be an editor.

There is worse that goes on in economics

If the above makes some editors in Biological Sciences look bad, then I apologise. Being an editor for a journal is a pretty thankless task and there is no financial gain, and precious little career gain to do an editorial stint. However, if you're going to do it, then you must do it well. The half measures that I describe above are simply not good enough. But biological journals are a huge cut above those in economics. I've always had my doubts about economics as a subject. Rather like theology, it's based on a fanciful construct that puts its own practitioners in positions of power when we'd do just as well to flip a coin.

In May this year I was contacted by the "International Journal of Finance and Economics" and asked to review a paper on ETF performance (that's Exchange Traded Funds, but I only knew that now because I looked it up). I deleted the email as I do get a lot of spam from predatory publications, although these usually ask for papers and not reviews. Later in June, I got another email again asking for a review.

 

I noticed that the journal was published by Wiley, and so I looked them up. It turns out that the International Journal of Finance and Economics is a real academic journal. So why were they asking me for a review on ETF performance? I wrote back: 

Hello. Please tell me why you chose me as a reviewer.

Mrs. Terry Wirtel wrote back with all honesty:

 

Looking at the website, I note that Mrs. Wirtel is not an Associate Editor, nor an Advisory Editor. It turns out that Mrs. Wirtel is actually the editor’s administrator. I then wrote to the editors Mark P. Taylor (Washington University in St Louis), Michael P. Dooley (University of California at Santa Cruz) and Keith Cuthbertson (Cass Business School) and explained that they should be doing their own editorial work, otherwise their publication is bogus. Perhaps not surprisingly, they didn’t write back.

Although somewhat amusing, this exchange is also a serious worry. When editors, like Mark P. Taylor (Washington University in St Louis), don't do their own work, they leave the reputation of their journal in tatters, and it is reduced to the equivalence of junk status. Interestingly, Mark P. Taylor (Washington University in St Louis) is also the editor of two more economics journals. I'd be surprised if he ever reads the content of any of them.

Summing up

The way to get round making the kind of editorial blunders I describe above is to read the manuscript. The guidance of how to read a manuscript should be explained to editors when they take up the position. There is plenty of information out there on the internet, but the journal’seditorial policy should be understood by all of the editors (and preferably open to authors and reviewers too), and that should include reading manuscripts before sending them out for peer review.