Weight loss jab concerns, and crafty bribing caterpillars
In this edition of The Naked Scientists: Weight loss jabs are now being prescribed by GPs in England. But do we know enough about the side effects? Also ahead: a new simple test that could safely replace endoscopies for patients with Barrett’s oesophagus. Plus, we’ll hear from one of the engineers at the Vera Rubin Observatory – which is shedding new light on the cosmos...
In this episode
01:06 - New risks associated with weight loss jabs
New risks associated with weight loss jabs
John Wilding, University of Liverpool
The weight-loss jab Mounjaro is being made available at GP clinics in England – but only for those who meet strict criteria. The weekly injection, which can help people lose up to 20% of their body weight, is already in high demand globally. But doctors warn supplies are limited, the drug is costly, and side effects like nausea and vomiting are common. There are also concerns it may interfere with how well the contraceptive pill works, and even rare cases of inflammation of the pancreas have been reported. I've been speaking with John Wilding, who leads clinical research into obesity, diabetes and endocrinology at the University of Liverpool…
John - So Mounjaro, which is also known by the name Tazepatide, is given by injection once a week. And the way it works is by mimicking two of the body's natural hormones that tell us after we've eaten a meal that our stomach is full and that we've had enough to eat. So these are hormones that are naturally released from the gut to tell us we're full. They also do other things like stimulate insulin secretion, which is why they've been used to treat people with type 2 diabetes. But it's the effects on the brain and on appetite control that are the reason why they're effective weight loss medications.
Chris - When people take them, they lose between 10 and 20 percent over time of their body weight. Why do they lose that? Is it just that they don't feel hungry so they eat less? Or do they eat less of the wrong things? Or is it a combination of those things? What do the trials look like?
John - That's a really interesting question. As you rightly point out, at the highest dose we see on average people lose about 20 percent of their body weight. What they're doing is partly mimicking these natural hormones so you feel fuller after you've eaten. The example that I often use is you eat a sandwich and you feel like you've had Christmas dinner. So it's more about making you feel full rather than making you feel less hungry. Although they do probably also help people to feel a bit less hungry as well. In terms of what people choose to eat, some people do tell us that they tend to crave less sweet foods and fatty foods and unhealthy foods when they're taking these medicines. But that evidence isn't really very strong at the moment.
Chris - Is there any evidence people also over time change their eating behaviour in the sense that when a person has had a course of this they reach closer to their ideal body weight? Do they then maintain that when they stop? Or are you condemned effectively to enriching the share price of the company that makes these drugs indefinitely in order to keep the weight off?
John - That's a really good question. The way I would frame this as a clinician who treats people with severe obesity and weight-related complications is firstly to say that of course as you lose weight eventually you will reach a plateau. Everybody will stop losing weight. Usually after about a year you reach a new steady weight while you're taking the medicine. If you stop the medicine then of course it stops working and at that point people tend to gradually put weight on again. Of course some people will be able to maintain the weight loss long-term. That's relatively uncommon. Most people would put the weight back on once they stop it. So we do need to think about these drugs much in the same way as we think about drugs for other long-term conditions like diabetes or hypertension or high cholesterol as treatments that people take long-term.
There is ongoing research that is asking the question whether it might be possible once you've reached that point of weight where you want to be whether you might be able to maintain the weight at a lower dose which would obviously be a good thing but we don't know whether that is the case yet.
Chris - What about the side effects? We've spoken to people who've taken these drugs. Sometimes they found that they had to stop taking them because they just couldn't handle those side effects. There isn't a free lunch with any drug, is there? There are always potentially side effects. What does the side effect profile for these agents look like?
John - You're absolutely right. Of course all medicines have side effects but what I would say is the most important side effect that everybody needs to know about is the fact that these drugs when you first start taking them particularly or when you increase the dose tend to cause nausea. Most of the time what we find is that if we increase the dose gradually that allows us to find the right dose that most people can tolerate.
Chris - And the impact on contraception? Because that's emerged more recently and I know a lot of GPs have expressed considerable alarm about that.
John - So in fact we've been using these medicines to treat diabetes for many years and as with all medicines when they're first developed we are very very careful in the clinical trials not to include women who might become pregnant because we don't know the long-term effects on a growing baby in the womb. So there is very little clinical data about the safety in pregnancy. There is a very specific issue with Tazepatide with Mounjaro which is as with all of these drugs it slows down the emptying of the stomach. So in one small study that was done in women taking oral contraceptives what we found was that it slows the absorption of the contraceptive and therefore that might potentially reduce the effectiveness of the contraceptive.
Chris - What does the economic argument look like? Because you've got politicians saying they want to do this on potentially a massive scale relatively speaking across a whole country in primary care so this would be family doctors who are able to administer these agents. Does this look like it makes for good value for money? Because they're not cheap these drugs.
John - So one of the jobs of NICE, National Institute for Health and Clinical Excellence is to actually look at not just the effectiveness of medicines but about the cost-effectiveness of medicines. And the NICE technology appraisal for Tazepatide has come to the conclusion that these medicines are cost-effective for use in the NHS and that is done for most medicines that we use in the NHS and this medicine comes into the same sort of range as many other treatments that we use in terms of its cost-effectiveness. One thing that the NHS has recognised is that to make that available to everybody who might be eligible immediately would be very expensive. And so what they've done is they've tried to target this at those people who are perhaps most likely to benefit early on. Now we can argue about whether the group that they have chosen is the right group. The rollout in primary care at the moment is only for people who have very severe obesity, a body mass index above 40 and have four out of five named complications from diabetes, heart disease, sleep apnea, high blood pressure and high cholesterol. So you have to have four out of those five and a body mass index above 40 because that's where they think the cost-effectiveness is likely to be greatest. That will gradually relax over time as we gain more experience and GPs become more experienced with prescribing the medicine.
Chris - The thing that I'm finding it hard to wrap my mind around though is that, as you've eloquently told us, this may translate into, say, a 20% reduction in body weight. But some people are 100% or 200% in some cases over their ideal body weight. So they're still going to be extremely overweight despite taking these drugs and running the risk of all the side effects that you've outlined. So is there a vanishing return there?
John - What I would say is that, of course, there are some people who are twice their body weight and if they lose 20% of their overall body weight, that's a substantial amount of the extra weight that they're carrying that they lose and about two-thirds of that weight that they lose will be fatty tissue, which is what we're trying to help reduce with obesity. Yes, of course, it's not going to get everybody to an ideal body weight, but there's no intervention that does that except for perhaps some of the most extreme forms of bariatric surgery that are very, very rarely performed because they have so many side effects. So we're actually, with these medicines, getting quite close to what we see with the more commonly used weight loss operations, like gastric bypass, and these medicines are a little bit less effective, but not that much less effective than these operations now.
09:50 - Capsule sponge test could replace invasive endoscopies
Capsule sponge test could replace invasive endoscopies
Rebecca Fitzgerald, University of Cambridge
Cambridge scientists have announced that a simple, quick, cheap capsule sponge test can safely replace expensive screening endoscopies – telescope tests – for around half of patients with the condition Barrett’s oesophagus. This is a change in the appearance of the cells lining the oesophagus close to the stomach. It’s relatively common, but it can lead to oesophageal cancer in some people, so patients with the condition require regular screening to spot people in whom this is happening. This is time-consuming, uncomfortable and does still miss cancer. But the capsule sponge test, where patients swallow a small sampling device on a thread which pops open in the stomach and, as it is withdrawn through the mouth, captures a large sample of cells on the way up, looks like a far superior way to do this monitoring, both in terms of practicality, cost and sensitivity. It’s the brainchild of pathologist Rebecca Fitzgerald, who’s just published in The Lancet the compelling results of a trial to assess the performance of her device in hundreds of patients across multiple UK hospitals…
Rebecca - Barrett's oesophagus is a response of the body to inflammation from acid and bile reflux. That's a very common problem, affecting about 10% of individuals. And in a proportion of those, actually in about 10% of those people with heartburn, that acid and bile damage can lead to the replacement of the cells in the lower part of the oesophagus. It's a kind of adaptive response with this condition called Barrett's oesophagus. Now most of the time, actually, that isn't diagnosed. So one thing we're very interested in is how we find people who've got that condition. And the importance of this condition is really about the risk of progressing to cancer. And this is a very horrid cancer. So finding it early is really, really important.
Chris - Is the problem then that we know that this happens, but what we don't know is in whom it's going to be a problem. But we have to look at everybody because if we don't look at everybody, because we don't know for whom it's going to be a problem, we're going to miss the ones that are going to turn into oesophageal cancer.
Rebecca - Absolutely right. So at the moment, we have a one size fits all approach. So when we find the Barrett's, we put everyone into a monitoring program. They come for endoscopies, depending a bit on the length of your Barrett's, but basically you'll be coming for endoscopy between every two, three, sometimes five years if it's a short segment. But most patients are going to be having multiple endoscopies over their lifetime. It's quite uncomfortable. It's quite stressful for the patient. It's expensive for the health service. And for most of those patients, they just don't really need that close monitoring. It's unnecessary. And that's what we think is exciting about the capsule sponge. It's very operator independent, usually given by a nurse. You know, it scrapes the cells out of the oesophagus when it's withdrawn in the same way, depending on, you know, regardless of who finishes your test. Unlike endoscopy, that really depends on how well the operator scrutinises the lining and takes your biopsies from the same place. And we can tailor it on your specific risk characteristics. So because of the markers that we're using, taking into account your age, your sex, the length of your Barrett's oesophagus. So that's all in our algorithm. We can then really start to kind of work at very low risk people who just frankly don't really need endoscopy. They could have a sponge every two or three years, maybe every five years when we get more confident in this. And then on the other hand, we really shine the light on people at the highest risk. And then they could really be sent to the most expert specialist centres with endoscopists that have got a special interest in this, who are going to use sophisticated imaging, couple that with really good pathology as well so that we find the early lesions and remove them and improve the outcomes. So it's a much more tailored approach, easier for the patients, more convenient, and we think it would be more robust.
Chris - The capsule sponge test is something you have pioneered. And I've seen it. And for the benefit of our listeners, imagine swallowing a pan scourer, but one that's tightly bound up in a small capsule, that then pops open in the stomach and can be pulled out. It basically captures some of the lining of the oesophagus on the way out, doesn't it? So you can look at that. So when you do this on people, how good is this as an intervention or a monitoring tool then?
Rebecca - Just to give your listeners an idea as well of how many cells we capture, 1 million plus, up to 4 million sometimes. So we get a much bigger cell sample than a pinch biopsy endoscopy. And then we do the tests, including the P53, the cell morphology, we couple it with AI in the study as well, to help the pathologists focus on the area of interest. And we have three risk groups, low risk, moderate risk and high risk. The low risk is over 50% of individuals, the chances of you having a serious problem in your oesophagus, if you were put in the low risk category, by that I mean what we call high grade dysplasia or cancer was 0.4%. Now the NHS trigger for a referral for endoscopy is a threshold of a 3% risk of cancer. So we're way below that. And that's how the study was designed to see where we were in relation to that 3% threshold.
Chris - Does that mean then, Rebecca, that what you're able to do is keep a lot of basically well people out of the clinic, you're minimising, but you can focus the endoscopies on the ones that it looks like they're the ones who are going to have the problem?
Rebecca - Absolutely right. So I think about 50% of people, we could just put them into a sponge-only follow up three years down the line, five years down the line, we want to do more work to see what that optimal interval would be. And maybe some of these people when we get more and more evidence could even be discharged. So that's over 50% of the people. On the other end of the spectrum, we have this low, moderate and high risk group. In the high risk group, when we saw our two key sponge markers as positive, that's a P53 test, it's a bit technical and the cell changes, which we call atypia. When you had both of those present on your sponge, the chances of you actually having something wrong at your endoscopy, so a dysplasia or cancer diagnosis is 85%. And that's only a small proportion of people who fall into that high risk category, less than 20%.
Chris - Does this mean then you have enough statistical power here to make the case and satisfy regulators, etc, that this should be the gold standard that we ought to be monitoring people in this way, in general practice in primary care? And is that what's going to become the established way of doing this now off the back of this Lancet paper?
Rebecca - This field has been stuck for a long time. And I think this level of evidence should enable us to start shifting. And it doesn't mean to say, of course, that we just kind of stop collecting the evidence. And we just say, that's great, we're going to change gear, you know, if you're in the low risk category, you're going to have sponge only monitoring, and we don't need to continue to evaluate that, of course, we continue to audit it. But I would hope we can start moving the bar for how we treat these patients, because at the moment, the system really doesn't work very well.
17:05 - Bribing caterpillar sparks evolutionary arms race
Bribing caterpillar sparks evolutionary arms race
Ritabrata Chowdhury, University of Cambridge
The jungles of Borneo is where the Macaranga tree grows. These are thick, bushy shrubs with large tasty-looking leaves. But they defend these leaves from anything tempted to take a bite using an army of ants that they billet all over the tree in special hollow structures in their stems. They’ve also evolved sharp hairy projections, called trichomes, that are there to snag stealthy caterpillars that know how to befriend, bribe and slip past the ant army. But researchers in Cambridge have discovered that one particularly crafty caterpillar species, Arhopala amphimuta, has taken the arms race a step further and, as well as sidestepping the ants, also evolved a much thicker skin to brush past the otherwise lethal hairs. Ritabrata Chowdhury made the discovery while on fieldwork for his PhD. We got together at Queens’ College, where he began by telling me what these caterpillars look like…
Ritabrata - They would be very difficult to find because they hide on the underside of leaves and they're very nicely camouflaged and their patterns match the veins of the leaves to an extent and very simply they look like a flattened grape. That's what somebody told me when I showed them this.
Chris - Are they grape-sized? How big are they?
Ritabrata - They are actually grape-sized. They're almost like two to five millimetres, not too big, and the largest ones can go up to like a couple of centimetres.
Chris - They do sound like very hungry caterpillars then if they're that big. But where do they live?
Ritabrata - What sorts of leaves do they prey on? These are very hungry caterpillars and they are mostly found in Southeast Asia. The ones which I studied, I found on the island of Borneo and these ones specifically feed on macaranga trees. These macaranga trees are defended by ants. So these are very stinging ants. Every day when I used to work there, I would get stung by at least 100 to 200 ants. Not very nice for me, but I think I have now become a bit immune to it.
Chris - When you say the trees are protected and defended by the ants, so the ants will build a nest near the tree and they swarm all over the tree and if anything tries to eat the tree or you come along, you get attacked?
Ritabrata - The plants actually host the ants inside the hollow stem. So these are mutualists or they are symbiotic ants. The ants live inside the hollow bark, basically the hollow stem of the tree and the plant provides it with food. And in turn, the ants protect the plant from herbivores of all sorts. And if anything tries to come along, even if it's a monkey or something tries to climb on the plant, they will sting them.
Chris - Or you.
Ritabrata - Or me, of course. Of course me.
Chris - And these caterpillars, you're saying that they have got some way of surmounting this or getting past these ants?
Ritabrata - Absolutely. So what these caterpillars do is that they bribe the ants with sugar-like solutions which they make from special organs called myrmecophilous organs and this is a very strong characteristic of the group of these caterpillars or butterflies. These are lysinids or lysinidae and they have this ancestral tendency of coexisting with ants where the ants do not harm them and in some cases, the ants even protect these caterpillars because think about this, no predator can get to them. The ants will sting them away too. And in some cases, these special aropala caterpillars, to deal with the ants, they actually mimic, they have the same chemicals on their body so the ants actually think that they are part of their own. They think they're an ant.
Chris - So not only do the caterpillars feed the ants so they're making them happy because they're stupefying them with a big meal but they also smell like they do. Is that the tactic?
Ritabrata - Yes, that is the tactic. This is one of the processes that these special group of caterpillars utilise to deal with ants and live on trees which are covered with ants and feed on them. And now comes the very hungry part because the ants don't attack them they can just happily munch on as much as they want.
Chris - How have the trees responded? Because the trees have already invoked the ants to defend them this fends off most hungry caterpillars that haven't got this tactic. The tree must presumably be fighting back in some way. Is there any kind of evidence that the tree is defending itself in an additional way against these caterpillars?
Ritabrata - In most cases no but this is where my story comes in. It turns out one plant has now because their primary defence which is in the form of ants has been kind of counter adapted or been dealt with they have now come up with a new technique which is this one single tree from Borneo, the macaranga tree from Borneo, seems to have come up with sharp hooked hairs on its entire surface. And when caterpillars try to walk on these surfaces these are specifically for the caterpillars that can deal with the ants when they try to walk on them they get pierced and they die.
Chris - And that's just one species of tree or one tree specifically?
Ritabrata - There's one species of tree. So the Macaranga trees there are around 300 odd Macaranga trees all around the world and they range from almost from Southeast Asia to Africa and out of more than 300 species only one seems to have evolved this hook trichomes which possibly which we thought might be to deal with these caterpillars and that's exactly what we tested.
Chris - How do you know that it evolved to deal with those caterpillars or could it be that actually just it's a self-fulfilling prophecy it has that trait for some other reason and it's also helpful for stopping these caterpillars causing a problem?
Ritabrata - The evidence which goes in the favour is that it's just one single species which has this hook trichomes. If you would expect we call them trichomes or the hairs because you would expect this to be a strategy for like generalist herbivores a lot of them would have it. So this seems like a de novo evolution especially because the ants have already been dealt with in this case. So of course we can't prove this at the moment but this is the most concrete hypothesis that we have against it given that it's just a single individual species in Borneo that has been able to do this.
Chris - So you've got this really interesting evolutionary story here. The trees evolve, the caterpillars come along and surmount the ant defence. So the tree then evolves these spikes which will hit the caterpillar where it hurts, and then the caterpillars evolve ways to get round the spikes.
Ritabrata - Yes, that's exactly the story that we're looking at. There's just one single caterpillar species that sems to use these plants, and they can walk on them.
Vera Rubin Observatory sees far and wide into the Universe
Freddy Munoz, Vera C. Rubin Observatory
The Vera Rubin Observatory has released its first breathtaking images: capturing galaxies, nebulae, and star clusters in stunning detail. In just ten hours of test operations, it’s already discovered over 2,000 previously undocumented asteroids, including several near-Earth objects. The observatory’s powerful wide-field camera is now set to scan the southern sky repeatedly over the next decade. Scientists say it marks the beginning of a new era in mapping our solar system and our wider universe. The summit mechanical engineer on the project, Freddy Munoz, was passing through Cambridge, so we got together for a chat, starting with why this telescope is so special…
Freddy - A very large field of view, and a very compact mechanical structure. So with those abilities, you can create at the same time, a telescope that can look very far away in light, which means very far away in time, with a lot of photons collection capability, going to very fine stars, and at the same time being able to move this telescope pretty, pretty fast. And that creates the ability to explore not only the distant universe, but actually the close near-Earth objects. You can conduct with this unique design, four different kinds of science with the same images, because you are collecting so much from the heavens, that you are creating a map of time, of asteroids, of galaxies, and the study of our local neighbour.
Chris - Is it just scanning the sky and collecting everything it can? Or do astronomers have to decide, this is our target that we're going to explore, and then you physically point it at something? How is it used?
Freddy - The telescope is driven by an automatic scheduler. So we have six optical filters, and a bit of infrared as well, that maps the sky following an automatic sequence of the entire southern sky, because this telescope is located in Chile, pointing to the southern sky. So the astronomer's role here is to define a sequence going over different patches of sky, and trying to avoid always the Moon, another very bright source of light, and the centre of the galaxy, elements that saturate the sensible electronic detectors. But at the very bottom, the telescope drives itself.
Chris - What do you do with all the data? Because presumably, there's a massive amount of information flowing off of these things, from these incredible, very, very deep, but also very wide images. So how do you process all that? What do you do with it?
Freddy - The data is the most significant part of this entire project. The telescope is just the instrument putting us in the flow of all this data stream. The facilities at the mountain collect all this electronic information in powerful servers that are then conducted to La Serena, Chile, the city nearest to the telescope. And then by optical fibre, it's transferred to three different facilities in the United States and in Europe, in order to create copies of this vast amount of data. At the same time, this software is collecting alerts, how every single object of the sky is changing night by night. The amount of data is the game-changing parameter for this telescope compared to its previous generation. In only a few weeks of work, the telescope has discovered more objects than other previous surveys have done in five years. And this is just the beginning.
Chris - Indeed, one of the reports I read said you got at least 2000 asteroids already. Were they missed because this has got amazing resolving power? Or were they missed because your ability to process the data is so extraordinary?
Freddy - To identify an asteroid, you need to take a picture and then compare it to another time, a minute after, 10 minutes after, hours after. And to do so, you need pretty much an automatic machine, which is the case that we have now.
Chris - So have you basically got supercomputers doing these image analyses for you? Because to churn through that much information and draw out those differences, that's not trivial computing power.
Freddy - The amount of IT technology, networks, cooling of servers, transferring of information between continents is unique. And it's the next generation of how science computing is driving astronomy. At the end of the 10 years survey, Vera Rubin Observatory will produce such a vast amount of information that will take decades for astronomers to process.
Chris - We got your work cut out for you there then. I mean, is it performing the way you expected? Were there any glitches? A bit like when Hubble went up into space, people then had a heart sink moment when they discovered, oh dear, there is a mirror problem, but they were able to fix it. Has it all gone to plan or are there issues?
Freddy - The challenges that we are experiencing now, it's the winter in Chile. We started operation of the telescope in very cold conditions as a starting point. And the main driver, it's this primary tertiary mirror safety. This telescope is capable of extraordinary speeds and accelerations, comparable to the speed of a people walking. But for a 400 tonne steel machine, that's a lot. So we are not using our full speed capacity yet. We are working towards to have this on full speed over the 2026 year.
Chris - And what are you hoping, if you had to sort of pick one thing and say, I really hope we're going to see that, what are you hoping this is going to reveal to you?
Freddy - This telescope will provide so immense amount of new objects that the astronomers will be overwhelmed for the next decade. It will reveal things that probably we don't know that it exists today. So this is not a project to confirm science, but to discover new things. I mean, we are excited to check what we're going to find in the next corner.
How long does the water cycle take to complete a circuit?
Thanks to Dr Jess Neumann for the answer!
James - How long does the water cycle take to complete a circuit from meltwater to evaporation, to snow and ice and back again? Over what size landmass and area does the water cycle take place? Thanks, Kate. All of the water on Earth, from what comes out of your tap to the water vapour in the air to sprawling polar glaciers made of ice is the same water that's been here for billions of years. The water cycle is the continuous flow of individual molecules around the whole Earth's atmosphere system, but we can also think of it on a more localised scale. Here's Dr Jess Newman, Associate Professor of Hydrology at the University of Reading.
Jess - If we take a little water droplet and it might fall straight into the sea, for example, it's come out of the clouds and precipitation straight into the sea, the sun comes out, heats the surface of the sea and we get evaporation of that water back up into the atmosphere. It could reasonably, within a few hours, fall back down as rainfall again. That's not impossible. So we could be thinking of it in terms of quite a localised area in terms of distance, but also in terms of time. And then we can have a case where perhaps this water, let's take from the cloud again, falling as rainfall and let's say it falls onto the surface of a leaf. It might then make its way, drip off the leaf onto the ground. It will then infiltrate into the soil. It can then make its way very slowly through the soil and the rock and may then go into something like a store, like an aquifer or groundwater. And potentially it could take years to reach that store and it could then stay in that store for hundreds if not thousands of years until it's obstructed back out or makes its way through what we call base flows or flow through the rocks back into our rivers. So it can happen pretty quickly and it can happen over really, really long timescales, depending on where that little molecule decides to go and which pathway it takes.
James - So the water cycle operates on many different scales, giving us many different timeframes for the journey of an individual water droplet. But let's zoom in on the particular section of the huge circular machine, so beautifully described by Jess, that we're interested in for your question, Kate. Melt water, liquid H2O released from warming snow and ice, usually from a glacier found predominantly in the polar regions.
Jess - So if we were to assume that that water gets released from its frozen state and it might've been in this frozen state in say a glacier or an ice sheet for thousands of years or hundreds of years, it might go straight into the sea. It can then stay in the sea as a storm or an ocean for a very long time, or it might evaporate and go up into the atmosphere. And generally the accepted, although disputed, decision about how long water stays in the atmosphere is eight to 10 days, but that's going to be different over different parts of the planet, depending on conditions and environments. But let's say that that ice sheet is on land or the glacier is on land. So that water droplet, again, it might go straight into the ground. So it becomes a liquid form and it goes, it infiltrates through the soil and into the rock. There's going to be loads of different pathways, almost infinite directions that this water molecule might take. And it's going to be very much dependent on what's happening in the landscape at that particular point in time as to where it's going to head.
James - It's fair to say, Kate, that you've drilled into some of the droplets with potentially the longest water cycles on earth, caught in several choke points, like being locked in an ice sheet or as a liquid in an aquifer deep underground. The thousands of years this might take to complete such a cycle, dwarfs the several hours some droplets may take to fall as rain, say, into a tropical body of water before taking off back into the atmosphere. Something which, for me at least, is fascinating to think about.
Comments
Add a comment