[smtf]

“Considering that spiders can already make really impressive geometric designs with their webs, it’s no surprise that they can take that leap to make an impressive design with debris and other things,”

Sure, but what is more interesting is how such a spider would know what itself looks like.

[pron]

I would recommend reading "Sex on Six Legs: Lessons on Life, Love, and Language from the Insect World" by Marlene Zuk. It's a truly wonderful book about insects (and spiders too), about their evolution, and about how little brain power is required to display complex behavior.

Here are a few of my favorite, more on the philosophical side, quotes from the book:

    Insects bring home the uneasy truth that you don't need a big brain to do big 
    things, and that in turn makes us question how the mind and, dare to say it, 
    the spirit, are related to the brain. It even makes us question what it means 
    to be human. What does it mean to have complex behavior? Does it mean you are 
    smart?


    Natural selection can produce what looks uncannily like intelligent thought or
    emotion but is no more than the relentless culling of minute variations in genetic
    makeup, generation after generation, for millions of years. Not only that, but
    insects too have small personalities, with some showing boldness in new situations
    and some hanging back with what looks an awful lot like shyness. It's turning out 
    that we haven't cornered the market on individuality, either.


    Insects are starting to answer the question of "What does it take?"—to have a
    personality, to learn, to teach others, to change the world around them—with the
    humbling and perplexing answer, "Not much." Humbling because they do these things
    with brains the size of a pinhead, and perplexing because if that's all it takes, 
    what does that mean for us, with our gigantic forebrains and exhaustingly long
    periods of childhood dependency?

[pagliara]

Thank you for the recommendation. You've inspired me to go read this right now.

[Someone]

It doesn't have to know what it looks like; it only needs to "know" what its predator's food looks like, or rather, what its predators are more likely to hunt.

And that probably is "know", as in "built into its genes through evolutionary trial and error". It is not likely this is something mothers learn their children, or that they adapt the looks of their decoys within days if researchers let new predators loose (either might be true, but would truly be surprising)

[6ren]

There's no need for the spider to know what itself looks like. If it builds something that increases survival/reproduction, that trait will be selected for.

  The spiders’ webs were crafted around face-height, near the trail

This makes me wonder if it was selected for stopping humans from wrecking their webs! It would explain why the decoys look like spiders, to humans.

[VMG]

way too little time to evolve that behavior

[brudgers]

Looking at the original article, it appears that the most spider like decoy photograph was chosen based on the appearance to humans - i.e. one with 8 legs which presumably you counted just as I did.

http://blog.perunature.com/2012/12/new-species-of-decoy-spid...

What appears spider-like to a fly, could be dramatically different - I'm imagining the process of trying to count appendages through a compound eye with a fly's brain and in the absence of the cultural system of symbols we use to represent numbers, etc.

See Thomas Nagel's What is it Like to Be a Bat? http://books.google.com/books?id=fBGPBRX3JsQC&pg=PA165#v...

[return0]

Flies are not predators for spiders.

[CamperBob2]

It would be interesting to amputate a leg from the spider in infancy and see if it creates a seven-legged simulacrum as an adult.

[jayflux]

Spiders legs grow back when you remove a leg from them.

[andrewflnr]

Really? I'm pretty sure that's not true. I think I would have heard about it when I went through one of my spider-fascination phases.

[robotresearcher]

It doesn't. It just likes to stick bits to its web in patterns. Feels good, (spider) man.

Something like this:

1. get bit of leaf. 2. go to center of web. 3. while standing on leaf bits, go down (with G). 4. drop bit of leaf, leaving gap to the left and right. 5. if more than 20 paces from the center. Forget it, relax. 6. goto 1.

With a bit of fiddling I have no doubt you can write a simple program to produce these shapes.

If you find this interesting, check YouTube for termite and wasp nest videos. Incredible!

[mirkules]

The process of producing decoy shapes isn't really a problem. Why these spiders made the shapes is more interesting.

I think that the spiders that didn't create these decoy structures were selected against in the evolutionary process. The production of decoys could have been genetically engrained into their instinct (via mutation), or, this species really are actually able to recognize itself and create a decoy -- both of these cases would have led to less of them being eaten by predators, hence natural selection.

I can't wait to hear the results of the study after they research it more closely!

[robotresearcher]

Which of these explanations seems more likely, in the context of a spider?

Spiders are very smart for their size (up to 80% of their body cavity is brain, and much larger than for insects), but giving them theory of mind is a stretch.

"That guy there wants to eat me. I look like this. If I make a thing that looks like me, he will think it's me and try to eat that, and I'll be safer."

vs. I just love to get bits of leaf and stick them just so!

[mirkules]

Sure, it's a stretch, I don't dispute that at all. Maybe a spider witnessed another spider being eaten, and since it was spared, decided to make another spider-looking thing so it doesn't get eaten again. It's not really a matter of self-recognition or theory of mind, but more of a pattern recognition, something spiders (I assume) are good at. "My friend was eaten, and I wasn't. I should make something that looks like my friend so I don't get eaten." But, then again, Occam's Razor and all that. :)

[Buttons840]

> "I should make something that looks like my friend."

I'm no expert, but the fact that spiders have 8 eyes suggests to me that they may, indeed, be good at recognizing visual patterns.

[takluyver]

The number of eyes isn't a particularly good guide. I'd guess most web-building spiders have quite poor detailed vision, as there's little need for it. Jumping spiders, which actively hunt prey, have good forward vision.

[bonobo]

I think that implying the spider is actually able to recognize itself is a little too far. It would require a level of awareness comparable to that of apes and dolphins. Not even dogs are able to recognize themselves when put in front of a mirror.

[sk5t]

While not a scientific article, here's an interesting read about corvids (crows, jays) presenting two sides of the argument that the birds possess theory of mind:

http://io9.com/5969515/corvids-the-birds-who-think-like-huma...

[ry0ohki]

A spider has to instinctively know what a spider looks like, otherwise how would it know what to mate with?

[VMG]

Knowing what it looks like doesn't mean knowing how to build it.

[Fargren]

Do the mate with each other or do they lay and impregnate eggs?

[nitrogen]

They mate somewhat directly: https://en.wikipedia.org/wiki/Spider#Reproduction_and_life_c...

[jre]

Some mussels have evolved a similar feature : they are able to make part of them look like a fish even if they are blind : http://whyevolutionistrue.wordpress.com/2011/03/02/amazing-m...

Probably those spiders are the result of a similar evolutionary process.

[mediocregopher]

For even more mind-blowing, a mussel that looks like a fish:

http://unionid.missouristate.edu/gallery/L_reeveiana/Reevian...

EDIT: looks like jre beat me to it

[seiji]

"know" is a bit of a strong word here. Does a printer "know" what a W is or does it just know where to put the dots?

[TeMPOraL]

Not that strong. For printer, the shape of a W is an external input. This spider outputs something resembling its own shape. So the question is here, why this and not something else (like randomly shaped clump), and how did it acquire this particular decoy blueprint.

[seiji]

Let me channel dawkins (and/or sagan) for a moment:

Imagine there's a cluster of spiders and a cluster of predators of those spiders. The predators can directly grab the spider, but if it fails and gets caught in the web, the spider wins. The predator is pattern matching on "what looks like a spider."

Maybe a few spider brains create reliable glitches in their webs. Maybe the glitches look enough like not-web where some predators get caught in the web when trying to catch the non-spider. The uneaten spiders-with-web-glitches reproduce and create further glitch making spiders. The predators not caught in the web (i.e. not fooled by the web glitch) also reproduce (n.b. the predators are reproducing glitch-avoiding pattern matching, so we have a glitch-vs-anti-glitch brain race), and some new spiders have more glitchy mutations making bigger anomalies. The anomalies looking more spider like will attract predators, and those spiders won't be eaten as easily. The uneaten spiders can reproduce more glitch-making spiders. The uneaten prey can reproduce more glitch-avoiding prey. The steady state of this model ends up with spiders reproducing themselves so the predator needs more and more advanced web glitch detection. Let this cycle go on for unbelievably long time scales (thousands? tens of thousands? hundreds of thousands? millions of years?), and you have a predator-prey derived "art."

The spider doesn't "know" it's intentionally making decoy spiders. It's just evaluating it's internal tiny spider brain state machine with evolutionarily derived glitches over the past ginormous number of years.

[mkeener]

Yes, Exactly! Another similar example would be the process that produced the heikegani crab.

TL;DR - Over years fishermen threw crabs with face-like patterns on their carapaces back into the ocean. Over time this produced crabs with extremely detailed faces, increasing the chances they wouldn't be eaten and could reproduce.

http://en.wikipedia.org/wiki/Heikegani

[btilly]

If you read your link, you'd find that Carl Sagan's theory is doubtful.

[bonobo]

We often hear about how long mammals' or birds' characteristics took to evolve, I'd like to have a grasp of the scale of time we're dealing with when talking about tiny creatures like these.

I have absolutely no idea of the time that is necessary for such characteristic to evolve, but I think it's not that much time when talking about insects or arachnids. Evolution is not iterated over time, it happens over generations. And when talking about insects or arachnids, we have a lot of them in a short span of time.

[hfsktr]

You raise an interesting topic I've never really given much thought to (time vs generations). Makes me think of other questions about genetics that I just don't know the answers to but that someone must have thought of.

Things like: How do they account for recessive traits? Can something that is recessive become dominant? Can something dominant just disappear (I imagine that it could with mutations happening so much)? Is there a specific animal they use to test these things? Do different traits take longer to evolve towards?

I imagine they use something that reproduces rapidly (fruit flies? fleas?).

On the topic of this spider and how does it know what it looks like...I don't know how much brainpower the spiders have but I know that we have a part of the brain that tracks where our body is (at least in relation to other body parts). Is it possible the spiders have something similar? Can this even be used to make a rough representation of oneself?

I am going to have to google some of these later as I'm now curious.

[glhaynes]

How about a printer with a picture of itself built-in as a test page pattern? :)

I'd guess that its sculptures started as fairly undefined clumps that served the purpose somewhat well (presumably that's what's built by other "sculpting spiders" that the article mentions?). And then evolution has refined that clump, like it does so well.

[ceejayoz]

Natural selection, over millions of years. Slightly more spider-like decoys get selected for.

[popopje]

because one with legs is a more effective decoy? (or whatever other implications a larger spider has)

[skc]

Having trouble viewing the video but what's blowing my mind right now is that in the picture I see, it has created a decoy with 8 legs. How does it know??

[robotresearcher]

The writer/editor chose that picture. Scroll down and there's a 5-legged decoy. It's still cool, but they chose the most impressive picture to lead with.

[skc]

Ah yes. Though I guess the spider could argue it isn't done sculpting yet.

[takluyver]

Looks like it. If one clicks through to the blog post, there are photos of two more decoys, both with <8 legs. But I guess most predators don't stop to count legs, so the evolutionary pressure for better decoys isn't that high.

[solox3]

More likely the case is it knows what its mother looks like.

[solox3]

What? it's not a "yo mama" joke. Many species of spiders show parental care: http://i.imgur.com/Ayrsv.jpg

[loopdoend]

So it turns out iPads make for dangerous projectiles. Let this be a fair warning to others.

[nitrogen]

Apparently spiders aren't the only ones with a long history of evolved behavior; before I even recognized what I was looking at, I had an immediate "WTF?!?!?!?!" fight or flight response to that fascinating image of a mother spider with many tiny spiders on its back.