‘Wadudu Castles’, or Furahan 'termite mounds'

I needed a background for a painting.

Every chapter in The Book at present starts with a double page showing a painting supposedly made by someone in the Furaha Universe. These paintings are all about humans on Furaha. Some are portraits of dignitaries, whose illustrious career earned them the privilege of forever having their countenance grace the Halls of Academia. Or, of course, until someone else comes along with even more vanity and money. Such portraits do not need much in the way of background; a generous helping of 'Van Dyke Brown' will do. But other paintings show Field Work, often in a much-romanticised version, and those are the ones that need a background showing some Furahan Biology. This one starts with wadudu. 

Click to enlarge; copyright Gert van Dijk

‘Wadudu’, as astute readers may remember, are small Furahan animals, somewhat resembling insects. If not, see here and here. Some wadudu may be communal species and build their own living arrangements, like termite mounds. Some early doodles of possible wadudu castles are shown above: these are ‘Mark I Castles’. The sketches show roof-like outcroppings or ridges along their surface, with a vague idea that they could protect the structure against the sun, or perhaps against rainfall or even snow. Termites are limited to hot regions, but there is no reason to apply that rule to wadudu castles too. 

Click to enlarge; copyright Gert van Dijk

I made a quick colour sketch of one design, shown above. This ‘Mark II Castle’ is not that different from a 'Mark I'; it has an elongated shape and is bilaterally symmetrical. The narrow side points towards the sun during the hottest times of the day, to limit the surface area that is directly insolated. The rootlike structures running along the ground are hollow cylinders that function as highways allowing the beasties inside easy access to wherever they need to go. Different castles may be connected by such cylinders, both on the ground and below it, allowing communities to link up and form even larger superorganisms. The overhanging roofs were smooth and white, to reflect sunlight and to protect the castle from heating up too much. 

 

Click to enlarge; From Claggett et al. J Struct Eng 2018; 144: 02518001; DOI: 10.1061/(ASCE)ST.1943-541X.0002043

The wadudu would have to produce different materials for that to work, so I had a look at Earth's termites. It turns out that termite mounds can have very different shapes (see above). The paper I copied the images from (Claggett et al DOI: 10.1061/(ASCE)ST.1943-541X.0002043) differentiated between conical, dome, cathedral, mushroom and meridian or compass mounds. The differences depend on differing circumstances and requirements to control temperature, humidity and ventilation. 

Click to enlarge; Oberst et al. J R Soc 2021; 18: 20200957;  https://doi.org/10.1098/rsif.2020.0957

It runs out that termites indeed alter the soil they use to build their mounds with. Termites use grains of sand or other materials and saliva to form 'boluses' that are then cemented together. They may use coarse granules in the centre of walls, and finer ones for wall surfaces. The termites can even mix clay in the boluses to provide water-resistant walls. With all this going on, it would not be a stretch to provide Furahan wadudu with equal or better capabilities to produce different materials, from watertight layers to highly reflective ones. That opens up some really interesting possibilities, so I think I will return to wadudu castles in the future.
 

The bilateral symmetry of the Mark II may not have been optimal. For most places on Earth, the sun will never be directly overhead. The axial tilt of Earth ensures that the sun will generally be towards the South if you are on the Northern hemisphere, and vice versa. This means that a sun shade does not have to be symmetrical in the North-South direction. The smart way would be build a shade that is angled in accordance with the direction of the sun, something that will depend on the latitude (and the rotation of the planet's axis). This is very similar to the way solar panels are positioned. Solar panels absorb light, but the shades should reflect as much of it as possible. They should be smooth and as reflective as possible; a glaring white would be nice. Depending on the climate, the nights might be cold, and in that case part of the castle could welcome the sun's rays early in the morning. That part could even be dark so the castle can warm up. The structure at the opposite end would catch the last rays of the setting sun, so you might think that the castle could do with some warming up before the night. Possibly; I assumed that the castle would warm up during the day regardless of the shades, so the roofs may need to protect against the evening sun too. The roofs could bend down towards the direction of the setting sun to do just that.  

Click to enlarge; copyright Gert van Dijk

Click to enlarge; copyright Gert van Dijk

I sculpted such a Mark III castle in ZBrush: here it is, shown from several points of view to highlight its asymmetrical shape. Instead of one roof, it has several. Imagine that it is about 3 meters high. The roofs should be white, whereas the rest of the structure will probably have the colour of the sand it is built from. 

The videos above show such a castle. Unfortunately, I could not colour it correctly for the video. I wished to show how it looks while the sun moves from sunrise to sunset in one smooth movement, but that did not work well. I had defined a few positions for the sun, but the software treated them like stops on a train ride: the sun now reaches a place, stops ('all aboard!') , and departs for the next place. I should do it again, but the point was to produce a painting, not a video. And I can now paint that background.
    
By now you will probably be wondering what the castle builders look like. Actually, I don't know yet. I haven't yet decided whether they the castles are built by wadudu, spidrids, or by another group altogether. Shouldn't I have decided that before I designed the castles? I think not: I doubt that anyone could deduce termite anatomy by studying a termite mound, other than that the builders are small, can move grains of sand around and stick them together. That'll do for now.

Zoom interview with Dougal Dixon tomorrow

 I received, courtesy of Dougal Dixon, an invitation from Oscar Salguero, a book curator from New York, to view a Zoom presentation given tomorrow by Dougal. Mr Salguero assured me anyone can register for free. Because time is short I will just pass on the information.

 

 "Mr. Dixon will give a live presentation highlighting some of his most important works on speculative zoology from the last 40 years. This will be presented via zoom and it's sponsored by the Center for Book Arts in NY."

 Friday, April 30, from 1 - 2:30pm EST

That should make 19:00 -21:30 hrs CET, and 18:00-20:30 hrs for people in the UK  

 Here is the event (click on "Register" button to receive the zoom link):

https://centerforbookarts.org/calendar/talk/speculative-zoology-fictitious-examples-of-factual-processes

Facebook page:

https://www.facebook.com/events/300204934901648

 ----------------------------------------------------------------------------------------------

 After the event

 This is just a short update for those who are curious about Dougal’s work.

In the talk, Dougal mentioned that there would be a 40-year anniversary edition of ‘After Man’, that should be published before the end of 2021. I understood that it will have more material than the books published so far, including sketches.

Another interesting bit of information was that there were Japanese model kits of some of the creatures featuring in ‘After Man’. Dougal showed a box with content of an unassembled kit of the Night Stalker. I was curious and wondered if I could find out more about those kits (secretly hoping I might still order one from some forgotten Japanese warehouse).

 

Well, I found two entries on a site of a company where you can indeed order such things, but the items were sold out. There were a few images, and I chose the one above, because it showed the contents of the box.

 

The two images above are were posted by a collector on a site called ‘dinotoyblog’. Interesting, aren't they? Alas, I found no hidden stores of such models...

 

Terra Ultima by Raoul Deleo

 A friend called my attention to an art project she had read about in her newspaper, and she thought I would like it. I certainly did and decided to write about it, even though it is not a typical speculative biology project. The project is much more about the art than about science, and has a strong poetic quality. But foremost, I think, it is work of extraordinary quality.

The project is called 'Terra Ultima'. There is a nice site showing quite a few paintings, as well as a Youtube channel, with some videos showing the artist at work. The background story is that the artist, Raoul Deleo, has visited the continent of Terra Ultima several times and came back with the sketches of its wildlife to prove it. 

Here is a nice video showing the background story. Raoul lives only about 35 km from where I live, has been working on his project since at least 2015, but I managed to remain completely ignorant of his work until recently. I was just in time though, as a book about the project has just been published. In case you are interested you should know that at present it is in Dutch only, but Raoul told me that a French version is already in the works, and English-language publishers have already showed an interest. Let me just jump in and give you some examples.          

Click to enlarge; copyright Raoul Deleo

The 'Pallid Octopossum' (Octopossum leucostolum). I chose to start with this one, because terrestrial cephalopods are a Big Thing in speculative evolution. A long time ago, I argued that the tentacles of walking octopuses (yes, I know about the plural 'octopodes', but this sounds better) would turn into legs with joints (here is the first post of a sequence; to find the rest, just type in 'walking with tentacles' in the blog's search window). The arms need not evolve a skeleton if they mostly have to deal with tensile rather than compressive forces, so a brachiating animal might still have tentacles (or a cernuating one). This cute fellow's furry tentacles would fit the bill.

Before anyone starts complaining that this animal cannot possibly have evolved either from a cephalopod or a mammal, I agree. The combination of fur, external ears, a long probably sticky tongue on the one hand and tentacles with suckers on the other tells us that this animal is a composite.

Regular readers may remember that I had my doubts about the 'mix a monster' approach used in the film and game industries (examples here and here). At the time, I concluded that that was one of the defining differences between the worlds of 'speculative biology' and 'creature design': the first frowns upon approaches that are contrary to science, whereas the second couldn't care less, as long as the results looks cool and sells. I try to limit my protesting against a lack of scientific rigour only when people make unjustified claims about scientific quality (I am a scientist; that is what we do!).

Click to enlarge; copyright Raoul Deleo

A gecko! Sort of anyway (Erminogecko mollidactylus). It is a hexapod! There's fur, external ears again, three pairs of legs with four toes each, so this animal is also a composite. Have a close look at its potential prey. In an interesting twist, this creature has one pair of membranous insect-like wings, as well as two bird-like legs, and camera eyes, not compound ones. It is almost the reverse of the geckoid, so the illustrator is letting the viewer see that he knows what he is doing, and that the mixture is part of the fun.

There is probably a third school of 'design an animal', and this particular one is probably much older than creature design or speculative evolution: it is 'animal fantasy'. I guess its ancestry includes the fantastical animals bored mediaeval monks drew in the margins of their books, as well as the beings Hieronymus Bosch painted to make people scared of Hell.

Click to enlarge; copyright Raoul Deleo

A 'lieveheersbeertje' (Coccinellursus hexapedus). A 'lieve-heers-beest-je' is literally a 'Good-Lord's- beast-diminutive suffix'. That's Dutch for ladybug. A 'beer' is a bear, so the name is a nice pun combining 'ladybug' with 'bear'. I wonder what the translators will do with that; do any readers wish to take up the challenge?

The animal is again a composite hexapod. Fantasy animals mostly have to look good, and does this one look good! I really must work on my own fur painting skills. Have you noticed that the animals are all posed in such a way that their anatomy is well visible, without awkward foreshortening and with good lighting, so there are no shadows obscuring part of their shape or colour? That style reminded me of James Audubon.

Click to enlarge; copyright Raoul Deleo

Here I put together Raoul's 'Flamingo fawn' (Phoenicopterus cervocephalus, which conforms to 'deer-headed flamingo') next to Audubon's flamingo. Look at how the head is turned around, so the shape of the animal just fills the available space. Raoul confirmed that Audubon was a great inspiration, so that explains part of the atmosphere of these paintings. 

 

Click to enlarge; copyright Raoul Deleo

The 'sheen green harefly' (Libellula lagoformosa). An insect-mammal composite showing a wonderful mixture of forms and colours. I cannot help wondering how it would fly, with those four wings at its very top. I guess it might look a bit like a tetrapter in flight, although the harefly obviously has bilateral symmetry. 

Click to enlarge; copyright Raoul Deleo

The 'headwinged penguin' (Pygoscelis cephalopterix). Humour is one of the best ways to avoid people taking your work overly seriously, and Raoul's work is suffused with a mild humour. Penguins look charming to begin with, and equipping penguin heads with functional (!) wings puts the work into the proper perspective, I think.

Terra Ultima is a charming and very entertaining animal fantasy. Even if you prefer a diet of rigorous scientific speculative evolution, the Terra Ultima site is definitely worth a visit, if only to be inspired by the extremely skilful paintings.     

I have already ordered the book, and will keep you informed on this blog when there are versions on other languages.     
   

Explaining and animating how cloakfish swim

In the two previous posts I wrote about parts of The Book that provide background information about how animals move on Furaha. There will be four double-page spreads showing such themes in The Book, about rusps, spidrids, tetrapters and cloakfish. 

I thought the cloakfish one would be easy, until I decided that it was high time I also made some progress towards a short CGI documentary I wrote about earlier, the one with cloakfish biodiversity as its main subject. That is a very big job and it is quite possible that I will fail on the programming side. But nothing ventured, nothing gained, so I went ahead and put some hours into Matlab programming. My strategy to design diverse cloakfish is to write editors that allow shapes to be designed with ease. The programmes then proceeds to make ‘3d meshes’, the basic working material for 3D design. How to get from meshes to nice photorealistic images is another story altogether. 

Click to enlarge; copyright Gert van Dijk

These are the editor screens. The user places control points here and there, which are then connected by a spline function. This results in nice smooth curves, useful for organic shapes. Panel A shows the body designer with a default shape. The inset shows a separate window controlling cross-sections of the beastie. By changing both shape and cross sections interesting forms can be produced. Panel B is the cloak editor. Apart from determining the shape of the cloak it also allows control over cloak movement, such as number of waves, wave amplitude, cloak curvature, thickness, etc. Panel C does something similar for the four front fins, and panel D shows the resulting output for the default shapes. 

Click to enlarge; copyright Gert van Dijk

With a few minutes’ worth of tinkering, you get this relatively slender cloakfish, probably a reasonably fast swimmer. 

Click to enlarge; copyright Gert van Dijk

Or this short and bulky ‘short-sleaved cloakfish’.

Click to enlarge; copyright Gert van Dijk

Or even this highly derived cloakfish, in which the cloaks are no longer ribbons, but shaped like penguin wings. They function in much the same way. 

 But the main point of this post was to show how the cloaks move. To do that, I had a look at the literature, and I found some papers on knifefish, but to my surprise more papers about artificial robot fish with similar fins. Apparently, people all over the world are working on robot fish, which is nice. Less nice is that some work at defence institutes, so what are they preparing for? Killer fish robots? Must we really? 

Click to enlarge; from: Liu, Curet. Swimming performance of a bio-inspired robotic vessel with undulating fin propulsion. Bioinspir. Biomim. 13 (2018) 056006 1748-3190/aacd26

Anyway, here is an example of robot ribbon fin design. The artificial cloak design followed exactly the same reasoning as my virtual cloakfish designs. Such papers make a distinction between ‘oscillation’ and ‘undulation’. If a cloak, ribbon or fin swings side-to-side as a whole, the word ‘oscillation’ is used, and when waves travel along the length of the cloak, it is ‘undulation’. But the distinction is not all that clear; it depends on the number of waves travelling along the cloak. If there is less than one wave on the cloak or fin at a time, then the movement largely concerns the fin as a whole, so the movement edges towards oscillation. Here is a YouTube film explaining the difference. 

Pure oscillatory sideways movement are useless, because they do not propel the beast forwards. The animation above shows such an almost pure sideways movement of the cloaks. The movement would push water away from the cloak, resulting in a force towards the attachment of the cloak (the ‘dagger’). In knifefish, with just one cloak underneath the body, this force ‘heaves’ the body up. With four cloaks, the dagger will not be going anywhere, so this is just a waste of energy. We want water to be pushed backwards. The cloaks push water backwards when its parts are at an angle to the direction of movement. These parts produce forwards thrust. 

Let’s equip our default cloakfish with exactly one wave per cloak. One half of that wave will be angled towards the left, and the other half towards the right. Both produce sideways forces, but these should cancel one another out. The robot designers reported that the robots swam nicely with just one wave per cloak. Mind you, the robots usually had just one cloak, like the knifefish. 

The animation also shows one other trick: if you look closely, you can see that the cloaks do not sway much at front, and the amplitude of the wave increases towards the back of the animal. I borrowed that from real biology, as at least rays and knifefish do this. 

Let’s now equip our knifefish with 2 waves per cloak. The parts of the wave that are useful for swimming are now at a steeper angle towards the direction of movement, nearing perpendicular to it. I thought that this should increase the propulsive force a lot, but work on the robot fish did not agree. The velocity did not increase much, but the robot was more stable, which is intriguing. Real knifefish have more than two fins on their ribbon fins at one time, so there must be an advantage in that. 

Click to enlarge; from: Blevins, Lauder. Rajiform locomotion: three-dimensional kinematics of the pectoral fin surface during swimming in the freshwater stingray Potamotrygon orbignyi.  The Journal of Experimental Biology 2012; 215, 3231-3241

This image is from a paper about ray fin movement. The fins as a whole move up and down (so they oscillate), while there are ripples along the edges of the fins (so they also undulate). Nature seems to like combinations better than separations. The edges of the fins curl up and down, so they do not move as if there are completely stiff rays in them. 

I decided to build that in too, so here is a ‘curly-cloaked cloakfish’. I like it. One odd thing about these cloakfish animations is that it is not immediately obvious how they work, when you see the movement. It is also not easy to find an angle, when you rotate the objects, from where it is easy to get an immediate overview just how the animal is built. That can be seen as a disadvantage, or, in reverse, as an advantage, because it underlines that we are looking at an alien shape. 

By now, cloakfish ‘evolution’ has progressed to allow a variety of body and cloak shapes. Shapes range from ‘long-sleeved’ cloakfish with long narrow ribbon fins with multiple waves along their surface, to very short star-shaped cloakfish with narrow wings that fly through the water. That is certainly enough material for two explanatory pages in The Book, and should be enough for a short documentary too. But that will depend on me improving my skills as regards merging and smoothing 3D meshes.

Explaining Spidrid walking

 This will be another fairly short post. In the previous post I explained that I was working on pages for The Book that explained some biomechanical tricks of Furahan lifeforms. Other such explanatory pages deal with things such as rusp snouts and gaits (done), photosynthetic spectra (done), cloakfish movements (to be done) and spidrid gaits (done).

I will not show the actual spidrid illustrations, but can show you some of the underlying thoughts by way of animations.  I used the programs -all Matlab- to produce such animations to choose a single frame, which I then rotated this way and that until I was pleased with the composition. The resulting image was then imported into Corel Painter and used as the guideline for a digital painting. But I will not show these here.

 

The first animation shows your typical run of the mill garden-variety spidrid. It is walking slowly, meaning that each leg is on the ground for one half of the walking cycle. With eight legs, it is easy to have enough legs on the ground to provide a stable support platform at all times. The ‘support diagram’ is a polygon connecting all feet that are on the ground at any one time.

The gait used here is what I call an ‘alternating ripple’. Imagine that the legs are numbered 1 to 8, going round the animal. If the phase differences are 1/8, 2/8, 3/8, up to 8/8 in the same leg order, then legs that are close in phase would also be close in space, so many legs on one side of the beast could be off the ground at the same time, so it would tipple over. So, we introduce an additional offset for even-numbered  legs: 1/8, 5/8, 2/8, 6/8, 3/8, 7/8, 4/8 and 8/8. You will probably need some graph paper to get to grips with all this...

The red lines show the path a legs traces in 3D space. Because the ‘camera’ follows the body, the tracer paths are also respective to the body. The effect is like that of the animal walking along on a treadmill.  

The second one is very similar, but the main difference is that the legs are on the ground for less than half the time. Such schemes are typical for fast movements. The animation runs at the same number of frames per second, so you cannot appreciate the speed difference. There are fewer legs on the ground at one time. The polygons of the support diagrams have now morphed into lines or even points, if there is only one leg on the ground.       

  

Now we move to a more specialised racing spidrid. The camera no longer moves along with the beastie, but is fixed in space. The animal has relatively long legs, and runs the risk of knocking them into one another. That has to be avoided by reducing stride length a bit and adapting the gait: adjacent legs should not be at opposite phases in the cycle, as they then will certainly knock into one another. This one is using the ‘slow’ leg pattern at which a foot is one the ground half the time.   

This is the same racing spidrid at high speed. Each leg is now on the ground for less than half the time, more suited to fast movements. The animal changed its gait in two ways: the first is that the phase differences between legs are smaller, and arranged in such a way that at times there is no leg on the ground at all. It is effectively jumping! The second change is that the leading leg could only contribute to forwards movement by powerful and fast flexion movement, and I decided that the flexion muscles are relatively weak, so the animal simply lifts that legs into the air. (It should really move the body up and down as well, but the animation was not designed for complete realism.)

Many of the principles here are quite common for earth animals: walking faster is often achieved by increasing cycle frequency, stride length, reducing the fraction of the cycle that a leg is on the ground and adapting gaits to achieve jumps. Most mammals and reptiles always use all four legs at all speeds, with a few exceptions (kangaroos, probably hadrosaurs; no doubt there are more). But these poor unfortunate creatures have to make do with only four legs anyway, leaving them little choice. Earth crabs do have choices, and when they speed up, they actually use fewer legs, down to just two. Spidrids, not to be outdone by Earth creatures, have similar tricks up their virtual sleeves.

Explaining tetrapter flight (Tetrapters/tetropters X)

Just a short post this time. 

The Book will not only contain paintings of animals, plants, mixotrophs and people, but will also contain explanatory diagrams. These are usually much more boring to produce than texts or paintings, but they still have to be done... I had postponed writing and illustrating the flight of tetrapters for quite some time, and have now decided to get to work and not to look up until it is done. 

The challenge here was how I could capture the complexity of tetrapter flight in static diagrams, although I already had videos dynamically showing how tetrapters move their wings. The two diagrams above form part of a set of eight. Together they depict one complete movement cycle. I decided that I would show the path of the tip of the wing in the diagram, and that a portion of the path would be shown with a bold line, to indicate the movement since the preceding diagram. I hope that works.

Click to enlarge; copyright Gert van Dijk

 

Click to enlarge; copyright Gert van Dijk

The two diagrams show the point in time where the wings are moving apart after the 'clap' phase, when they touch or nearly touch. When they then 'fling' away, they create the 'clap and fling' mechanism that provides part of the lift. For more on that, you may read some older posts indicated by their year of publication: 2009a, 2009b, 2011 and 2018. 

 

Click to enlarge; copyright Gert van Dijk

 By the way, I have started to update the main Furaha website. I will gradually add some new material, but do not want to give away too much of the content of The Book. Still, some newer paintings will creep in here and there. I changed the image on the welcome screen, and do not think I ever published that particular rusp image before. So there you are.

The evolutionary origin of dragon flight

Perhaps some of you know the ‘dragon’ series of books by Marie Brennan. They are very enjoyable fantasy novels, describing the consecutive adventures of a woman in a fictional world resembling late 19-century Earth quite a bit. There are of course disparities, such as the existence of dragons. 

The protagonist is a young woman, or she is at the start of the first book; in later volumes she is older. She wishes to study Natural History, not at all a proper endeavour for a young woman in her rather Victorian world. To find out more about her and her adventures, you should read the books for yourself, because this post mostly deals with the cover art of one the books: 'Within the sanctuary of wings'. 

You can have a look at all the covers on the site of the artist, Todd Lockwood, where you can also buy prints and have a better look at the covers. 

 

Click to enlarge; copyright Todd Lockwood

Here is the cover in question. I think you can guess why it caught my interest, well, other than this being creative and high-quality art. The four creatures form an obvious succession, starting with a six-legged dog-lizard wingless beastie in front. The distance between its hind and middle legs is larger than between the middle and front legs, which is interesting, but the really important detail is the little membrane connecting the hind part of the middle leg to the torso. 

In the next stage, that membrane is a lot larger, and that second 'protodraco' seems to be gliding as much as it is running. The middle legs stick out sideways a bit; do they even hit the ground? The third stage has undeniable wings, and the fourth stage is all about wings. That one is a proper dragon, with four walking legs and two wings. I think this cover is a diagram of the evolutionary history of flight in dragons. 

The covers of all these books betray an obvious biological interest, which is completely in keeping with the content of the books: the heroine builds her career on the scientific study of dragons, and the readers get to hear interesting snippets here and there. When I saw the cover, I was curious whether this apparent evolutionary history also featured on the books, but I did not find that there. 

Would this progression from a running hexapod with six walking legs to a clade with four walking legs and two wings work? Regular readers will know that Furahan hexapods gave rise to flying forms as well, so I had to devise an evolutionary background for them, just as had been done for these dragons. Actually, hexapods did not only gave rise to the four-winged tetrapterates, but also to two-winged bipterates. On paper there is even a third flying clade, but I haven't worked much on that one yet. 

A basic concept in evolution is that an organ or feature will not develop just because some future descendent, a few million times removed, may make good use of whatever the organ does. That structure must convey some advantage in its early stage, or why would it be there? That advantage is unlikely to be true flight. But the future wing could help in gliding down from a tree or a cliff, or it could help the animal to jump higher and longer, or to run better. If you read about the evolution of flight in pterosaurs, birds and bats, you will find such explanations. I think that the winglets in the 'protodracones' on the cover show that, in dragons, flight evolved 'from the ground up', starting with more efficient running. 

Click to enlarge; copyright Gert van Dijk

At present, that is not how Furahan tetrapterates or bipterates got into the air. They took the 'down towards the ground' route. The rough sketch above was my first try at finding a shape for such animals. This is an arboreal species with folding leg flaps on all six of its legs. They crawl up into trees and then jump out of them and glide away, somewhat clumsily. 

Once the evolution of such gliders is jump-started (sorry for that one), the next stage is likely to involve optimization. The animals will probably not be very good at flying in the beginning, so it may pay for them to have an aerodynamically stable body plan. A stable flight scheme would make them less maneuverable, but at least it should keep them airborne without sophisticated neural control. The easiest way to achieve a stable shape would be to have the centre of lift directly above the centre of gravity, and in turn the easiest way to get there is probably to use the middle limbs as the main or only wings. 

There are intriguing other possibilities though, because the hexapod Bauplan in principle allows  not just one, but two or even all three pairs of limbs to be used for flight. With less than three pairs the next question is which pairs of limbs should be used for that? Any such design must also look into how the remaining legs must be adapted, so the animal will make aerodynamic sense while also being able to move about on the ground. 

Such things might be good material for another post, but I would not be surprised if readers run away with these and better ideas long before I will ever get around to writing that post.