The Planet That Keeps Shrinking plus the Weekend Wrap

0:00Hello and welcome to Astronomy Daily. It's Saturday the 12th of September, 2026. This is Series 5, episode 192. And this is the Weekend Wrap. I'm Anna. And I'm Avery. Anna, our lead today is a planet getting smaller, which is not a sentence I expected to say this week. Mercury. And not smaller as in a revised measurement of what it is now. Smaller as in how much it has physically lost since it formed. The planet has been contracting for 4.5 billion years as its interior cools. And it writes the evidence on its own surface. A new paper says we've been reading that evidence wrong in one very specific and rather beautiful way. By up to 30%. The old figure for how much Mercury's diameter has shrunk was something like 4 to 16 kilometers. The new one goes as high as 23. And the reason we missed it is that the thing

1:02doing the hiding is the same thing that has been resurfacing Mercury for 4 billion years. Craters. Craters. We'll take it properly because the method is as interesting as the number. And because there's a spacecraft arriving at Mercury in about 10 weeks built to settle it. Then the week that was, and it was full. A rocket reaching orbit from Western European soil for the first time. The moon possibly assembled in five hours. 27 new worlds beyond Neptune. 109 radio bursts weighing the universe's missing gas. And magnetars turning out to be half of everything rather than one in a hundred. Plus one we didn't get to during the week. New evidence that our own sun is capable of a super flare. And two fresh developments from the last 48 hours. Rocket Lab has gone to the government accountability office over that 700 million dollar Mars contract. And Starship's first orbital flight has moved again.

2:03And the sky for both hemispheres, which this week has a supernova in it you can go and find yourself. Let's get into it. Let's kick things off with Mercury, shall we? Start me with the basic physics. Why would a planet shrink at all? Because it was born hot and it has been losing that heat ever since. Mercury is a small planet with an enormous iron core. Around 85% of the planet's radius, which is wildly out of proportion compared with Earth. Hot rock and hot metal occupy more volume than cold rock and cold metal. So as the interior cools, the inside of the planet contracts. And the rigid outer shell has to accommodate a smaller interior. And a solid shell can't just deflate smoothly. It can't. It has to go somewhere. And it does that by breaking. The crust gets pushed together, thrust faults form, and one slab of crust rides up over another. On the surface that shows up as a cliff,

3:04long, sinuous, lobe fronted, sometimes a kilometer or two high and hundreds of kilometers long. They're called low bait scarps. There are wrinkle ridges and high relief ridges too. And the whole family goes by a wonderfully plain name, shortening structures. Because they record the surface getting shorter. And Mercury's the textbook case for the solar system. We've known since Mariner 10, flew past in 1974 and came back with images of these things everywhere. Discovery Roops is the famous one, a scarp 500 kilometers long cutting straight through craters and offsetting their rims. So how do you turn cliffs into a number? Very directly. Every thrust fault has taken up a certain amount of horizontal shortening. And you can estimate it from the height of the scarp and the angle the fault dips at. Map every shortening structure on the planet, add up all the shortening, and that tells you how much the circumference has reduced.

4:04Divide through and you get the change in diameter. And that's where the old number came from. Roughly four to 16 kilometers off the diameter. And it had a problem everybody in the field knew about. It was lower than the physics wanted. Model Mercury's thermal history and the models predict more contraction than the surface appears to show. A mismatch with the surface on the small side. Which usually means one of two things. Either the model is wrong or you're not seeing all the evidence. This new work argues it's the second. For a reason that's almost embarrassingly simple once somebody says it out loud. This is Gaku Nishiyama, a planetary scientist at the German Aerospace Center in Berlin with colleagues in Japan, published in geophysical research letters, and released by the American Geophysical Union on Thursday the 10th. And the simple reason is impact craters bury the cliffs. Every asteroid that is hit Mercury over four billion years has thrown out a blanket of pulverized rock. And that debris drapes over

5:09whatever was there before. A sharp kilometer high cliff gets softened, partly filled, buried at one end, broken into pieces that no longer read as one continuous structure. And then you sitting at a desk mapping images either don't see it or map it as something smaller than it was. Is that a hunch or did they measure it? They measured it. And that's what makes the paper convincing rather than merely plausible. They used messenger. NASA's Mercury Orbiter, which went round the planet 4,105 times between 2011 and 2015, before being deliberately crashed into the surface. It carried a laser altimeter and a dual imaging system. The team did two things with that. First, stereo photogrammetry. Take two images of the same ground from different angles, and the parallax gives you topography. It's the trick your two eyes play to give you depth perception, applied to a planet. So three dimensional terrain where before there were flat pictures.

6:11At much better resolution than the altimeter alone, and across parts of the planet, the altimeter never properly covered. And second, the clever bit. They measured surface roughness independently and asked a question. Does the density of shortening structures you can see depend on how rough the surrounding terrain is? And it does. Strongly. The rougher the terrain, which is to say the more heavily battered by impacts, the fewer shortening structures you find per unit area. Now, there's no physical reason the interior should have contracted less under rough ground than under smooth ground. The faults don't know what the surface looks like. So that correlation isn't geology. It's a detection limit. It's the signature of evidence being erased, and you can use the strength of the correlation to estimate how much has been erased. It does indeed, which is the best thing a bias can do. Nishiyama's own analogy is freshly laid

7:14gravel hiding the ruts in a road. The ruts are still there. You just can't see them from a moving car. So what's the corrected number? Between 10 and 30 percent more contraction than previously estimated. In round figures, the total loss of diameter goes from a range topping out around 16 kilometers to as much as 23, about 14 and a half miles, call it an extra seven kilometers hiding under rubble, which doesn't sound enormous for a whole planet. It doesn't, and Mercury is about 4900 kilometers across, so we're talking a fraction of a percent. But the number isn't interesting because it's big. It's interesting because of what it constrains. Nishiyama's line on that is the one to hold on to. He says, more shrinking means Mercury could have a larger metal core or fewer light elements like silicon mixed into that core, or a higher starting temperature. Unpack that. Why does total contraction tell you about the core?

8:15Because the amount of planet trinks is a thermometer reading integrated over its whole history. Iron contracts as it cools and contracts again when it solidifies. So if Mercury lost more volume than we thought, either there was more iron to lose it from or the core was pure. Light elements like silicon or sulfur change how iron behaves as it freezes, or the planet simply started hotter and had further to fall. And the mismatch with the thermal models closes, which is the quietly satisfying part. Nishiyama says the corrected amount actually makes sense to him, meaning the surface record and the predictions now agree rather than pulling against each other. There's a pattern here I want to name because we hit it twice already this week. Wednesday the galaxy m 74 found to be more than twice its catalog size because the catalog size was really a statement about how deep the survey went. And yesterday magnetars turning out to be

9:17half of all neutron stars because the catalog counted how long each kind stays visible rather than how many are born. And today a planet that's shrunk by more than the map says because the map is drawn on a surface that's been partly erasing itself for four billion years. Same lesson three times in one week and it's the most useful habit of mind in the field. Before you ask what the universe is doing, ask what your instrument and your sample are letting you see. Which brings us to the spacecraft that's about to do it again properly. Beppy Columbo and the timing is genuinely lovely. The joint European and Japanese mission has been flying since 2018 and after nine planetary flybyes it's now in the arrival phase. It separated its big electric transfer module on the third of September which we covered at the time. Gravity capture at Mercury is the 21st of November so about ten weeks away. The Japanese orbiter Mi-O is released around the 9th or 10th of December.

10:18The European orbiter reaches its final science orbit on the 10th of March and routine science begins on the 6th of April. And what does it bring to this specific problem? Two things messenger could not. The laser altimeter, Bella, is substantially more capable with vertical precision quoted down to the tens of centimeters. And orbit geometry, which matters just as much and gets mentioned less. Messenger flew a highly eccentric orbit so it had superb resolution over the northern hemisphere and much poorer coverage of the south. Beppy Columbo's European orbiter sits on a far less eccentric polar orbit, even coverage of the whole planet at consistent resolution. So the small structures that were being missed get counted. And that's a proper falsifiable prediction out of this paper. Which is what you want. If the shortfall really is buried small structures, Beppy Columbo should find a population of modest scarps and ridges in exactly the

11:18rough terrain where current maps look suspiciously empty. If it looks at that ground at 20 centimeter precision and finds nothing, the correction is wrong and the thermomodels have a real problem. Either way we'll know within a couple of years. And there's an Australian thread in this one. There is and its infrastructure rather than science. Which I think makes it better rather than worse. Everything Beppy Columbo does at Mercury has to come home through a dish. And one of the dishes is in Western Australia. Esa's new Norse station, about 140 kilometers north of Perth and run in partnership with CSIRO is where Europe's deep space network began. The 35 meter antenna there was the agency's first. There's now a second 35 meter dish at the site built for the current generation of missions with Beppy Columbo named among those it supports. Same reason the Canberra complex exists. Same reason and its simple geometry. A spacecraft is only visible from part of

12:19the Earth at a time. So continuous contact needs dishes spread around the planet in longitude. Which means southern stations and is also why ESA built one at Malargue in Argentina. The southern hemisphere isn't a nice to have in deep space communications. It's load bearing. When the first detailed topography of Mercury's southern hemisphere comes down next year, some of it will have arrived via a paddock in Western Australia. One last thing. Mercury is actually in the sky this week. Barely. Very low in the Western twilight. Setting inside 40 minutes of the sun and a difficult catch for everybody. We'll come back to it. But I like the symmetry. The hardest planet to see is also the one whose surface has been hardest to read. And for the same underlying reason. Something keeps getting in the way. Right. Let's move on to the week that was. Five stories from the week day run in the order they happened. One we didn't get to. And two fresh developments from the last day or so.

13:21And the theme of the week, if it had one, was honest revision. Almost everything on this list is somebody finding out that a number we were comfortable with was wrong. Let's start at the beginning of the week Monday and the launch story of the week on Friday the 5th at 12 minutes past 10 in the evening local time. A rocket called spectrum lifted off from Andoya spaceport in northern Norway and reached orbit. Built by a unit company, Ezar Aerospace. And it's the first vehicle ever to reach orbit from Western European soil with a real payload, not a mass simulator. Five university Cube sets Berlin, Trondheim, Meribor, Vienna and the Bulgarian company in Duroset. Plus a fixed experiment from de cubed. It went into a stretched orbit and circularized on a second stage restart which is non trivial on your second ever flight. The first in March last year failed about 30

14:24seconds in on event valve. And the thing to watch now is cadence, not the milestone. Exactly. Vehicles 3 through 7 are in production and the Munich factory is built for more than 30 a year. Reaching orbit once is a headline. Reaching it on schedule is a launch industry. We paired it with Gilmore Space in Queensland as the southern hemisphere version of the same ambition. Eris flew 14 seconds from Boeing last year and Test Flight 2 is now listed for early 2027. Next up was our big moon story for the week. Tuesday and the headline everybody else ran was that the moon formed in five hours. Which isn't quite what the paper said and the difference matters. It's a sensitivity result. It's a sensitivity result. Kagan Denton and Robin Canop at the Southwest Research Institute with Eric Asfog in Arizona

15:25in astrophysical journal letters. The canonical story is a Mars-sized body called Thia hitting the proto-earth, throwing a disc of debris into orbit, and the moon slowly accreting out of it. What this group added was something nobody had properly included, rock strength that changes with temperature. Because previous simulations treated the rock as a fluid. Essentially, and rock isn't a fluid, it has strength, and that strength collapses as it heats. Put it in and for some impact conditions, you skip the disc entirely. The collision directly produces a single intact satellite in about five hours. Not always, a hot young Thia, under 60 million years old, gives the immediate moon. A cooler, older one gives the classical slow disc. So the finding is that the outcome is sensitive to a parameter we'd been ignoring. A genuinely important result and a much less exciting

16:30sentence. Follow-ups are the full parameter survey and deep lunar samples. The two roots give different interiors, so there's a test. We closed it on the Jack Hill Zircons in Western Australia. The oldest bits of earth anybody has held. Moving on to Wednesday, 27 new worlds passed Neptune. Wednesday's lead was Hubble and Webb working the same patch of sky together, and pulling out 27 previously unknown trans-Neptunian objects. The faintest ever directly detected, the smallest around five kilometers across. Five times below what ground-based surveys reach. And the surprise wasn't the number. It was the colors, two papers in the astronomical journal on the eighth. Morgan at Northern Arizona on color, Eduardo at Victoria on the size distribution. The expectation was that small objects are collision fragments, so they should look like

17:31Hubble, mixed, homogenized, no memory of origin. Instead, the small ones keep the same color relationship as the large ones in both the dynamically cold population and the hot one. David Trillings' line was that the hot objects retain a signature of where they were born. Which points back to how planetesimals formed in the first place. It supports rapid formation directly at large sizes rather than slow grinding up from dust with arokoff as the type specimen. And the next act is Rubin at Sarropachone in Chile, which will find these in bulk, plus occultation chasing from Australia and New Zealand. Now, I know this next one is a favorite of yours, from Thursday, weighing the universe with radio bursts. Thursday, and yes, it is indeed one of my favorites of the year. 109 localized fast radio bursts, mostly from the deep synoptic array at Owen's Valley, used to measure something nobody

18:34could pin down properly before. How far galactic feedback has pushed gas out of galaxies and smooth the clumpiness of matter in the universe. And the mechanism is the nicest thing about it. A fast radio burst is a millisecond flash, and every free electron between us and it slows the low frequencies slightly more than the high ones. So the burst arrives smeared, a chirp, and the size of the smear counts the electrons along the line of sight. Critty Sharma, Vikram Ravi, Elizabeth Kraus, and colleagues, nature astronomy on the eighth. A prism made out of the entire intervening universe. And the result. Gas fractions in big halos running about 1.9 sigma above stacked Eroceda X-ray measurements. The bursts count cool gas the X-rays miss, and clustering variants cut by something like a factor of eight at the scales where feedback bites, which bears directly on the S-8 tension.

19:39The southern spine of that story is long. The first fast radio burst came out of parks, Murrayang in 2007, and the McCourt relation came from ASCAP in Western Australia. Moving on to Friday's episode, magnetars are half of everything. Yesterday, the one that still feels too big. Magnetars, neutron stars with magnetic fields around a hundred trillion times Earths have always been the exotics. About 30 confirmed against several thousand radiopulsars. One in a hundred give or take. And the new number is one in two. Roughly one in two at birth. Selsa Pardo Arauho and Nandareya in Barcelona with Michelle Ronkey and Vanessa Graber, nature astronomy on the 10th. A population synthesis modeling every class of isolated neutron star as one family, evolving spin down, magnetic and thermal decay and galactic dynamics together.

20:42Among the 24 known neutron stars younger than 2000 years, magnetars and central compact objects are about 59 percent. And the catalog was never counting births. It was counting visibility. A radiopulsar beams for tens of millions of years. A magnetar burns bright and fades fast. Count sightings and you count lifetimes, not births. Two consequences. The galactic supernova rate has to go up to about two per century, double the long standing figure. And magnetar central engine models for superluminous supernovae gamma ray burst platose and fast radio bursts suddenly become affordable because there are enough engines to go round. Which is a direct handshake with Thursday's lead. Within a day of each other from opposite ends. And the southern thread is foundational. The whole field starts with SGR 0526-66 in the large

21:48Magellanic Cloud in 1979. And the modern end runs through the Merchison Widefield Array in Western Australia. Now I believe you also have a story that we ran out of time to run during the week. Indeed, the one we didn't run during the week and it's a proper story. On Thursday, the Max Planck Institute for Solar System Research with the University of Colorado published New Evidence that our sun is capable of a super flare. Define super flare because the word gets thrown around. A flare, an order of magnitude or more beyond the biggest our instruments have recorded. The kind of energy release that makes the Carrington event of 1859 look like a warm-up. Two years ago, the same institute surveyed more than 56,000 sun-like stars in Kepler data and found stars like ours appear to produce super flares roughly once a century each. Which was uncomfortable because we have four centuries of sunspot records and no super flare in them.

22:54So either we're unusual or we're overdue or the comparison is wrong. And this paper goes at it from our own star rather than from other stars. Natalie Kravova and colleagues took the 300 strongest solar flares recorded between 2010 and 2016 and correlated the energy released in each with the size of the active region it came from. You get a scaling relation, bigger magnetic region, more available energy. And then you extrapolate it to the biggest spot we've ever seen. A sunspot group from 1947, the largest in 400 years of systematic observation, covering about 6 tenths of 1% of the solar disk. Run it through the relation and a region that size holds enough stored magnetic energy to power a super flare. Kravova's line is blunt. Our sun has super flare potential. It can produce massive

23:55sunspots and those can serve as the starting point for the most extreme bursts of radiation. Kaviyat's please, this is the kind of result that gets a terrible headline. Three and they matter. An extrapolated scaling relation is not a prediction. It says the energy budget is there, not that the sun will spend it. 1947 produced large flares but nothing like a super flare, so having the potential is demonstrably not the same as using it. And the spot coverage on the genuinely superflaring Kepler stars is still well beyond anything our sun has shown. What's the independent evidence it has ever happened? That comes from Earth and it's rather wonderful. Tree rings and polar ice record sudden spikes in cosmic genic isotopes. Carbon 14 in wood, burrillium 10 in ice. And there are several sharp events in the record, the best known around

24:56774 of the common era and another around 993. Almost certainly extreme solar particle events far larger than anything in the instrumental era. So the geological answer is yes, it happens on roughly millennial timescales. And the practical read? Not alarm, infrastructure. A keratin class event today is a grid and satellite problem and that's exactly what space weather forecasting exists for. Australia's own space weather forecasting center sits inside the Bureau of Meteorology and issues the warnings, airlines, power operators and satellite fleets act on. It's also the reason the Aurora Australia gets spectacular. The same particles, a much smaller dose. And before we get into skywatch, I believe you have a couple of story updates for us. Two quick developments from the last couple of days. The first is a sequel to a story we

26:00opened on the third. NASA awarded Blue Origin a firm fixed price contract worth about $700 million to build the Mars Telecommunications Network, a relay orbiter on a blue ring bus launching on New Glenn, delivered by the end of 2028 and operational at Mars by 2030. Rocket Lab was the losing bidder. And Rocket Lab has now protested. Filed with the government accountability office on Friday 11th, two grounds that the award appears inconsistent with the eligibility criteria, Congress mandated for the procurement and that NASA's technical review of Rocket Lab's proposal contained what the company calls incorrect assertions and conclusions. Their framing is that procurement standards exist to ensure fair competition and protect public investment. How do these usually go? Historically not well for the protestor. And there's a neat irony in the

27:03precedent. When Blue Origin protested NASA's Lunar Lander Award in 2021, the GAO denied it that July. Blue Origin sued and lost in November. And NASA awarded Blue Origin a second lander contract anyway in 2023. The GAO works to a statutory 100 day clock, which puts a decision around mid-December. And we should say plainly that neither NASA nor Blue Origin has responded publicly as we record. They haven't, and we're not going to guess at the merits. This is a procurement dispute between two serious companies and will report what the GAO finds. And the second update? Starship Flight 14, which we've tracked since booster 21's static fire cleared at the end of August, has moved again. It was no earlier than the 15th of September. As of Thursday the 10th, it's no earlier than the 18th. Same vehicles? Booster 21 and ship 41, both flying for the first time,

28:08both Block 3 from Pad 2 at Starbase. Third flight of Starship Version 3, and the headline objectives are big. The first attempt at an actual orbital trajectory, and the first deployment of real satellites, around 20 operational Starlink V3 spacecraft. And I want to correct something we said on air earlier in the month. You do, and I'm glad you're doing it. When we first previewed this flight, we described it as including the first attempt to catch the ship itself with the tower arms. The current public flight plan has that catch deferred to a later mission. The booster is targeting a water landing in the Gulf, and the ship a splash down in the Indian Ocean. So, orbital trajectory, Starlink deployment, water recoveries. No tower catch of the ship on this one, as things stand. As things stand, and SpaceX has a habit of changing the profile late, so we'll take it as it comes. No earlier than the 18th. And to the sky for the week ahead,

29:14a good one because the moon stays out of the way. New moon was yesterday afternoon, so we're into thin evening crescents, building to first quarter on Friday the 18th, and dark mornings all week. From Sydney and similar latitudes, Venus rewards being prompt. Low in the west after sunset, unmissable at magnitude minus 4.8, heading for greatest brilliancy on Friday the 18th. Note that date. Some listings give the 22nd from a different definition of the peak. We use the 18th. And there's a conjunction right on top of us tomorrow and Monday evening. A very thin crescent sweeps past Venus, half a degree apart at closest, a moon's width in front of Spica in Virgo. One of the lovely naked eye sites of the year, and Southern observers get the better geometry. The pair sits higher at the same stage of twilight than from North America. If you own a camera and a tripod, Sunday evening is the one. And Mercury, since we spent 10 minutes on it.

30:20Worth trying and be realistic. Magnitude minus 0.5, which is bright, but only two degrees up 20 minutes after sunset. You need a flat western horizon, clean air, and binoculars to find it before your eye does. A tick the box observation rather than a spectacle. But there's something to be said for looking at the planet we've just spent a segment taking apart. Saturn? The week's reliable telescope target for everybody, building towards opposition on the 4th of October with the rings about 7 degrees open. From the south, it rises in the east in the evening, and rides high through the middle of the night. Rings plus Titan is a five minute look that never gets old. North America, your turn. Saturn's the same target, different timing, up around midnight and about 50 degrees high by two in the morning, which is superb altitude for detail. Two satellite events in the small hours for telescope owners, Dione transit Saturn's North Polar Region from about 255 eastern for

31:22roughly 40 minutes, and Tethis slides into Saturn's shadow around 10 past two. And Mars is your pre-dawn object, up about half past one and working through Gemini. On the 18th, it passes six degrees south of Pollix, an easy color comparison of orange planet against orange giant. Now the one I'm most pleased about, the Supernova. This is the week's observing gift. There's a Type 1A Supernova going off in the Galaxy NGC 7331 in Pegasus. SN 2026 double AIV picked up by the Atlas survey at the start of the month and sitting around magnitude 12, peaking near the 10th. That's comfortably within reach of an 8 inch telescope under a decent sky and well within reach of a modest camera on a tracking mount. The Galaxy's a lovely target in its own right. A bright spiral about 40 million light years off often called the Dineb Galaxy with the dear lit group in the same field.

32:22Sources differ on the distance anywhere from 30 to 45 million light years, so treat 40 as a round figure. North America has the clear advantage. Pegasus is high overhead in your evening, close to ideal. From Sydney, it's a real challenge. The Galaxy sits at about 34 degrees north declination, so from 34 degrees south, it only reaches around 22 degrees above the northern horizon, through a lot of atmosphere and usually a lot of city light. Doable from a dark site with a clear northern aspect, around 10 to 11 in the evening. And why it's worth the trouble? Because a Type 1A is the standard candle the entire accelerating universe result is built on. The thing three Nobel laureates were defending at the end of August. And three days ago, we covered Chandra finding 84 hypersoft x-ray sources that may be the progenitor systems producing exactly this kind of explosion.

33:22So when you put an eyepiece on that faint dot in Pegasus, you're looking at one member of the population that measures the expansion of the universe. Not a bad Saturday night. Zodiacal light as well. And the equinox rule applies. So it's a both hemispheres item with opposite instructions. We're inside two weeks of the equinox on the 22nd and the ecliptic stands steeply to the horizon, which is what makes this faint cone of dusk-added sunlight visible at all. From the south, it's an evening object. West, after full darkness, a tall faint wedge rising from where the sunset, the false dusk. From the north, it's the mirror image, pre-dawn in the east, the false dawn. Either way, dark sight, no moon, patience. This new moon window is the best chance until early October. Safety passage. Yes, and it's in every episode for a reason. With Venus, this bright, some of you will try to find it in daylight. And it is a real observation. Venus

34:25at minus 4.8 is visible in a blue sky if you know exactly where to look. Do not sweep the sky near the sun with binoculars or a telescope to hunt for it. And do not try for mercury in twilight with the sun still up. Concentrated sunlight through any optic causes permanent retinal damage in a fraction of a second, with no pain to warn you. If you're ever looking at or near the sun, use a filter certified to ISO 12312-2. Fitted over the front of the instrument, never at the IP send, and inspect it for scratches or pinholes every single time before it goes near your eye. Looking further ahead. Two for the diary. Saturn at opposition on the 4th of October. And on the 6th of October, a pre-dawn lunar occultation of Jupiter. The moon, passing directly in front of the planet, build as the year's spectacular event. We'll build a proper curtain razor near the time. And that's the weekend wrap for Saturday 12th of September. Mercury has

35:26lost more of itself than we thought, as much as 23 kilometers off its diameter. And we missed it because four billion years of impacts have been quietly burying the evidence. Beppy Colombo arrives in about 10 weeks with the instrument to check. Looking back on the week, a rocket reached orbit from Western European soil for the first time. The moon may have assembled in five hours rather than centuries. Hubble and Webb found 27 new worlds beyond Neptune. 109 radio bursts weighed the universe's missing gas, and magnetars turned out to be half of all neutron stars. Plus, new evidence our own sun has the magnetic energy budget for a super flare. Rocket Lab has taken NASA to the GAO over the Mars relay contract, and Starships' first orbital attempt is now no earlier than the 18th.

36:27And the contact form on the site is real, and we do read it. More than one story this fortnight started as a listener question. If there's something you want us to take apart properly, tell us. We're back on Monday with the weekday run. I'm Anna. And I'm Avery, clear skies, and if you're in the southern hemisphere, go out tomorrow evening and look west. The moon and Venus have a degree apart in front of Spica. You won't need a telescope, and you won't forget it. The star is the toe. The star is the toe.