How do spiders know how to build webs?

Factual Questions
21

Richard Dawkins, in his book River Out of Eden, talks about how complex instinctive behaviors can arise in relatively simple organisms. In the book, he posits how several complex insect behaviors might have arisen, particularly convincingly, IMHO, with respect to honeybee figure-eight “wiggle dances.”

He also gives examples of how these complex behaviors can be short-circuited in unexpected circumstances. For example, when a bee dies, it emits a small amount of oleic acid. If a live honeybee is painted with a drop of oleic acid, it will be pushed out of the hive by the other bees, despite its protests and despite the bee being very much alive.

Also from the wiki entry, “a turkey will kill anything which moves unless it cries like a baby turkey. If the turkey is deaf, it will mercilessly kill its own babies.”

22

[QUOTE=robby]
In the book, he posits how several complex insect behaviors might have arisen, particularly convincingly, IMHO, with respect to honeybee figure-eight “wiggle dances.”
[/quote]
I read recently that honeybee dances do not encode directions to the source of the nectar, contrary to long-held beliefs. I’ll look for it, but if anyone else read it and could provide a link I’d appreciate it.

I’m predicting a Monty Python quote pretty soon in this thread.

23

[QUOTE=Lemur866]
One conceptual problem you might be having is the idea of a “blueprint”. It is highly unlikely that spiders have an innate blueprint for building a web, or birds have an innate blueprint for building a nest.

They don’t have a blueprint, they have a “recipe”. A set of stereotyped behaviors that, in the right environment, will result in a nest or web of a certain type being built. The spider doesn’t have a concept of a finished web in its mind when it is building the web. Rather it has a certain set of instructions…perhaps something like:

  1. Attach silk to a branch.
  2. Jump off and wave around until you reach another branch.
  3. Attach silk to branch, repeat until there is some signal that satiates the behavior.
  4. At this point the spider has the “outline” of the web, an irregular polygon.
  5. Then it climbs to the top and drops down to the bottom.
  6. Then start weaving lines from the first crossline, each line a certain body length apart from each other. Stop when you get back to your first line.
  7. At this point you’ve got the “spokes” of the web.
  8. Then start adding the sticky webbing in a spiral from the center, stop when you reach the outer polygon.

Of course, how these “steps” are entered into the spider’s firmware nervous system is unknown, but we can observe the methods animals use to construct structures. And they might use completely different rules for web building than I outlined above, but slight changes in the rules could result in very different finished webs, and very different starting evironments could do the same.

For another example, beaver dams. A beaver doesn’t have a plan of a finished dam, it has a simple rule…when you hear running water, add more sticks and mud. Any time the dam starts to leak, the beaver hears the water trickle and gets the feeling it should pile more sticks on the dam. But if you play a tape recording of trickling water, the beavers will keep adding sticks until the dam is enourmously overbuilt. And of course, we still don’t know on a physical or neurological level how or why the sound of trickling water makes a beaver want to cut down a tree and pile up the sticks, but it’s a lot easier to imagine how this behavior could evolve from simpler behavior. Likely the ancestors of dam-building beavers built lodges out of sticks, and the sound of trickling water meant their home was leaking, and time to reinforce it. And a slight change in this behavior lead to the creation of dams.

So if you think “recipe”, or a set of behavoirs that act together to create a structure, rather than “blueprint”, animal construction behavior won’t seem so baffling. Not that it’s explained, exactly, but just that the behavior doesn’t come out of nowhere.
[/QUOTE]
\

This is a good point. I assume nature found a relatively optimized way to transmit the information. Possibly the spider just has a few simple rules (not intended to even be a good guess, just an example):

  1. Hang web in a circle like motion.
  2. The distance between threads should be about x.
  3. Stop after y circles.

The thing that amazes me is the amount of information that would need to be coded into DNA (body and brain structures, instincts, etc.). Are there enough pairs to code everything that is coded? Are there other chemicals involved that carry information that we are not giving credit? Is the information not really at a gene level? Maybe one pair is logically part of multiple information carrying structures simultaneously?

24

[QUOTE=Sage Rat]
If you don’t accept it as simple instinct, encoded and passed on as genes, I’m not sure what else there is to suggest instead. Magic? // it’s still got to come down to something passed on through a similar process as passing on genes.
[/QUOTE]
Well, first of all, this isn’t about what I am and am not willing to accept, which is neither here nor there. (Personally, I’m ready and willing to accept anything if there’s good evidence to suggest it’s the right thing to accept.)

What this is about is the factual answer to what I think is a fascinating question, and how much we know about it. I am quite willing to believe that the behaviour is somehow conveyed genetically. My problem is that I have never yet come across an explanation as to how this particular type of information is, or can be, conveyed, given what we know about genetics. I’m not saying such an explanation doesn’t exist – just that I haven’t come across it yet. Nor has anyone else I’ve yet discussed this matter with.

I’m not a fundie, I don’t have a problem with evolutionary theory, and I’m not trying to make a case for an explanation involving either magic, myth or faith. It’s just that as far as I know we have at best an extremely limited understanding of how behaviours such as web-building are inherited, and I thought I’d see if the Teeming Millions could provide some illumination.

I happened to choose spiders and webs because I like spiders and I’m fascinated by them, and I think web-building is a particularly interesting and challenging example, more so than birds building nests or beavers making dams.

25

[QUOTE=CurtC]
I read recently that honeybee dances do not encode directions to the source of the nectar, contrary to long-held beliefs. I’ll look for it, but if anyone else read it and could provide a link I’d appreciate it.
[/QUOTE]
I’d be interested to see this, because I thought it had been demonstrated quite conclusively that the waggle dance did convey information regarding the direction (relative to the sun) of food sources - the experiments I saw involved a computer controlled robotic bee, programmed to dance in a specific way and resulting in statistically significant numbers of bees flying off in the expected direction afterwards.

26

How do you know how to breathe? Behavior can be “just there” in the genetic design of an organism. (I don’t see why voluntary vs involuntary muscle movements would matter.)

27

[QUOTE=Lemur866]
For another example, beaver dams. A beaver doesn’t have a plan of a finished dam, it has a simple rule…when you hear running water, add more sticks and mud. Any time the dam starts to leak, the beaver hears the water trickle and gets the feeling it should pile more sticks on the dam. But if you play a tape recording of trickling water, the beavers will keep adding sticks until the dam is enourmously overbuilt.
[/QUOTE]

Is this true, or just an example made up to convey a point? Could you point me towards a cite because this seems like a pretty cool example.

To the OP, apologies that this board is being rather prickly, but you’re certainly right to have stumbled upon a very complex question in neurobiology.

Here’s one example that I think might help you to get a better grasp on this from the level you seem to be coming from: how do cells know what they should “grow up” to be?

That generic, globular “cell” that all of the 4th grade text books have doesn’t really exist. Rather, every cell turns from a type of stem cell into a more specialized cell like a muscle fiber or a neuron. The way that this occurs is incredibly complex and basically cells have ways of “listening” to their surroundings to figure out how many of other type of cells there are out there and develop into the appropriate cell lineage. Again, very complex and we’re just beginning to scratch the surface, but here’s an example of a stem cell (a cell from a small embryo before those cells begin to develop into special tissues, etc.) that scientists have been able to grow into healthy pancratic cells (it’s when these cells go bad that diabetes develops).
http://www.nature.com/nbt/journal/v24/n11/abs/nbt1259.html

Basically, they bathe the cells in a specific line of chemicals and they get that result. Certainly, a similar process takes place in your brain and it develops its own complex anatomy.

So, to address your issue directly, why couldn’t that physical programming apply to neural connections as well to program innate behavior?

28

This is a fascinating thread. Thank you, Ianzin.

Thank you Colibri for your great explanation, and reason for lack of available explanation. It is awesome what we still need to understand. I have had the good fortune to observe spiders weaving webs closely. This discussion is revolving around the most advanced web - the total orb web of the garden orb weavers and others who do a full orb. It is by far the most logical to use, but that leads me to ask if Lemur866 is observing a different starting behaviour.

[QUOTE=Lemur866]

  1. Jump off and wave around until you reach another branch.

[/QUOTE]

Those I have observed release silk into the breeze (which needs only be a very slight to catch it), until it makes contact with something on the other side. Then they race across and adjust that thread to make the initial strong cross thread. Do you know of species which jump off and wave around as described above?

I’m not being picky - I do realise you were giving an algorithm for example, rather than a scientifically pedantic response. I was wondering if there were species which did this.

Our garden orb weavers, in Australia, usually leave that thread when they take down the web in the morning and reuse it the following night if possible. If it is a totally still night, and they don;t have a trhead already there, I have watched spiders give up on getting that thread out, and not build a web for that night.

I have also watched successive generations of our black house spider - which started with a single female inside the shed where I work, and so I know almost certainly that the first young was the offspring of the mother I had been ‘talking to’ for two years. She never did respond, but I yakked away anyway. The mother died after the birth of her young.

The mother and child constructed their webs in the same place - these are cribellate webs - woolly not sticky - which get new parts each night constantly growing. They built in the corner of the window, which got very hot in the middle of summer, and both then went and made a new web behind a rug hanging from the window sill.

The intriguing thing was that the mother was very confident- she would stay out at night on the edge of her funnel in the web when I went over for a chat. Her child, even after over a year in residence, would not stay out. If I approached, she retreated. A third black house spider who took up residence would not even stay out when I turned the light on. Three spiders, same species, probably a mother and two offspring, and three consistently different behaviours.

I believe this is evidence of personality, and there is no reason to think spiders wouldn’t have individual personalities just as other animals do.

So we are back to how much of personality is a response to environment and how much is genetically coded - as was raised in the twin studies.

Spiders are incredible!

Lynne

29

This may be my only chance to say ‘Hi Opal!’ without getting flamed for it, so I’m going to use it. Hi Opal! and thanks for playing. I may be completely wrong about this, and merely blinded by my own ignorance, but it seems to me that there’s a difference between the example you provide (‘knowing how to breathe’) and similar, and the building of a spider’s web.

Part of this difference lies in the fact that building a web is a construction project that has to be completed over a period of time, and a very complex one.

Building a web is a construction project on a scale far greater than the size of the spider. It involves (AFAIK) at least four distinct construction stages (initial lines; spokes; radial lines; finishing off, tautening, checking and optional addition of stickiness layer). These stages have to be completed in order. None serves any purpose except as part of a successfully completed web (ie all would otherwise be a huge waste of the spider’s energy and resources).

Whatever the underlying instincts or neuro-biological rules that the spider is using / obeying, the ‘algorithm’ (using the word in a loose and colloquial way) has to be sophisticated enough to cope with whatever circumstances the web-building spider happens to find itself in. Even for one given web-spinning species in one region of the world, the spider might have to cope with a huge range of prevailing weather conditions, winds, available sites and possible anchor points. Building a web on rigid anchor points in an airless site (eg my old bicycle frame in a closed garage) is quite a distinct task from building one between a tree and a gorse bush while both are swaying in the night-time breeze.

Note also that each part of the web’s construction has to have the ‘correct’ spatial relationship to other parts for the web to be successful, even though (as noted previously) the construction is on a scale sometimes scores of times greater than the size of the animal itself. And remember that the animal has (in most cases) virtually no vision beyond ‘moving gray shapes immediately ahead’, and certainly no way of ‘visualising’ what the web looks like to us, with our privileged ‘overhead view’.

I think this degree of specialisation and complexity marks web-building as worthy of special consideration. I think it is a far more complex behaviour than either nest-building or dam-building, which seem to me to call for less complex ‘algorithms’ (again, using the word colloquially). I think that web-building presents or offers a special challenge if we want to account for it successfully in terms of hard-wired instinctive ability that is communicated genetically. I’m happy for this to be the explanation. I just think we’ve a long way to go before we can say we really understand it.

30

I don’t see that there is any clear line where simple behaviour ends and complex behaviour begins; all living things have behaviours/actions built into them - some of these are conspicuously a mechanical function of construction, others are meta-actions built up from the interaction of many different sub-actions. Complexity is what makes it difficult to study and therefore a bit of a mystery at present - in much the same way that I can easily comprehend the way a transistor works, but find it difficult to fully comprehend and appreciate the totality of action performed by many tens of thousands of transistors acting together in the CPU of the computer from which I am posting this.

31

On the off chance you’re invoking a sort of irreducible complexity argument, here, I’ll point out that you can make a functional web by following the simple algorithm of “string strands of silk up every which way”, and presumably the first web-spinners followed just such an algorithm. As the spiders evolved, so did their web-building behaviour, and the webs became more efficient in the process.

32

Here’s a couple of websites that repeat the story I told about beavers responding to sound, although I haven’t found any research papers online to substantiate it.

http://www.naturealmanac.com/archive/beaver_dams/beaver_dams.html
http://wlapwww.gov.bc.ca/vir/pa/Beaver-Guide.pdf

As for the steps taken by spiders to construct webs, those were purely made up.

Now, back to Ianzin’s further questions. It isn’t true that “partial” webs are useless. You’re surely familiar with the webs known as “cobwebs”, right? These are webs built by spider species that are constructed more or less at random. The spiders just run sticky threads every which way and build up a mass of threads. These simple web types are probably the ancestral web type, and the more elaborate orb webs evolved from simpler webs.

And you’re correct that the spider can’t see the completed web, which is why the notion of a blueprint is misleading. The spider isn’t following a plan, it’s performing behaviors that, if conditions are right, will result in the finished web.

One interesting thing about animal behavior, especially non-mammal behavior, is how easy it is to “break” the behavior. Put an egg under a bird and the bird will treat the egg as it’s own egg and care for whatever hatches out of the egg. Even though cuckoos exploit this behavioral rigidity. I’ve put duck eggs under hens, and the hens take care of the baby ducks like their own chicks, although they get very upset when the ducklings go swimming. Or you have a situation where baby birds peck at a spot on their parents beak to stimulate regurgitation of food. But the baby birds will peck at a larger spot in preference to the parental spot.

So animal behaviors often only work in the environment they are adapted to. If the animal finds itself in an unexpected environment, the behaviors won’t work. So an attempt to build a web in an unsuitable environment just doesn’t work. So my suspicion is that these web building algorithms don’t have as much flexibility as you might think…it’s just that you don’t see many failed attempts because those spiders don’t live very long.

33

[QUOTE=Lemur866]
As for the steps taken by spiders to construct webs, those were purely made up.
[/QUOTE]
Can you please clarify what you mean by this? If you are referring to the stages that I described earlier, I’m at a loss to understand what you mean by saying they were ‘made up’.

34

When I described a stepwise algorithm for building a web, I just imagined one that could possibly work. Lynne pointed out that spiders actually use a different method to construct their outer guy lines.

But there really are plenty of spiders who don’t create the classic orb webs, there are lots that make simpler webs that don’t have that complex structure. So the complex webs didn’t have to evolve at one step, they started from simple webs that became more complex via small steps.

35

I realize there is a vast difference in sophistication between breathing and building a web. My point was merely that our genetic material already has built into it directions for “activities” rather than physical structures. Thus, it doesn’t seem difficult for me to take that a step further and say that some species’ genetic information includes behaviors as complex as web building.

36

[QUOTE=Colibri]
Spider web-building behavior, like most other behavior of arthopods and most other invertebrates, is innate and hard-wired. Each species has their characteristic behaviors that results in the web typical of that species. Spiders do not learn this from their parents or any other spiders.

How exactly this happens is one of the great mysteries of biology. We do know that many behaviors are genetic; crosses between individuals from species or populations with different behaviors can result in offspring with intermediate behaviors. We don’t at this point have a good idea of how this happens.
[/QUOTE]
To further clarify (or obscure, depending on your point of view) Darwinian natural selection posits that innate characteristics which promote reproductive fitness are inhereted, but acquired characteristics (which are learned or developed by a creature during its lifetime) are not. Lamarck proposed the opposite; that a blacksmith’s child would have big forearms because his father developed them performing his trade. Knowing, as we do now, of Mendelian genetics, we realize that this is a fundamental absurdity, but in the mid- to late-19th Century, this was not at all clear.

The ability to build a web is a benefit to the spider, hence it evolved progressively complex abilities to build and maintain webs. Since we have a very limited fossil history of the development of spiders we can’t really say precisely how this ability came about and the timeline for development, but there are clearly different levels of capability and complexity between different species of spiders, from those which can’t spin webs at all to those that make the elaborate spiralling webs across your front porch.

Web building, by the way, is really not that complex of an activity. You can easily write a program in, say, LOGO or Lisp (or if you want to endure real pain, C++ or Java) that will make a web-shaped result. You can program it to make the lateral members red (sticky) and the radial members blue (not sticky) just like a real web. It’s not hard to imagine that a neural complex sophisticated enough to handle respiration and do even basic visual and auditory recognition could have web-making algorithms to make a web.

Stranger

37

I remembered years ago reading about the effects of different drugs on spider’s innate web building abilities. Here’s a site that deals with it (wikipedia has a section too):

http://www.trinity.edu/jdunn/spiderdrugs.htm

Not entirely convinced the webs are that different. The LSD one just seems a little edited to make it look mildly trippy.

38

[QUOTE=lynne-42]
Those I have observed release silk into the breeze (which needs only be a very slight to catch it), until it makes contact with something on the other side. Then they race across and adjust that thread to make the initial strong cross thread.

…

Spiders are incredible!

Lynne
[/QUOTE]

In fact, here is a link that has (at the bottom) a nice diagram of the construction steps involved in one type of orb web: here.

I thought that I remembered reading somewhere that the 3D funnel webs were more complicated to construct than an orb web - but now I can’t quickly find a reference. (I’ll look more later.)

Fascinating question - but I’ll bet you could ask it in a similar vein: when an object approaches our eyeball (external stimulus as input) why do we blink and/or flinch involuntarily (behavior as response)? We don’t know how it is genetically transmitted, but that is a behavior that is hardwired in us. Another, now that I think about it, is shutting our eyes during a sneeze. So behavior can be carried in our genetic code - to build complicated patterns of behavior, link a series of simpler behaviors together (as Lemur866 and Colibri suggested).

39

[QUOTE=Stranger On A Train]
Web building, by the way, is really not that complex of an activity. You can easily write a program in, say, LOGO or Lisp (or if you want to endure real pain, C++ or Java) that will make a web-shaped result.
[/QUOTE]

A quick Google turns up, for example: Analysing Spider Web-building Behaviour with Rule-based Simulations and Genetic Algorithms and then a Q&A with one of the authors.

40

[QUOTE=Stranger On A Train]
Web building, by the way, is really not that complex of an activity. You can easily write a program in, say, LOGO or Lisp (or if you want to endure real pain, C++ or Java) that will make a web-shaped result.
[/quote]
Yes, it is easy to write a program that will produce a web-shaped result. Many people can do it. But it is not easy to write a program that will run inside a machine that will produce an actual web structure in the real world, with next to no visual pattern recognition, given arbitrary starting conditions and possibly difficult weather conditions. This is very different, and no-one has been able to do it.

I don’t think it was your intention, SOAT, to trivialise the ‘program’ needed inside the spider’s head. But I do think that this and other responses along the lines of ‘hey, it’s really not all that complicated’ run the danger of seriously under-estimating the complexity of the task and the corresponding complexity of the ‘program’ that has had to evolve. It is not just a case of ‘go round and round and lay equi-distant threads until you find it’s the middle’. The evolved ‘program’ has to accommodate a huge range of possible problems, hindrances, prevailing conditions, situations, locations and events that can occur during web-building.

[QUOTE=Stranger On A Train]
It’s not hard to imagine that a neural complex sophisticated enough to handle respiration and do even basic visual and auditory recognition could have web-making algorithms to make a web.
[/QUOTE]
Point taken and I agree. That having been said, I’m not aware of any spider with auditory recognition, even at a ‘basic’ level. And the visual apparatus of most of the spiders I’m referring to (AFAIK) is very poor. Vision only plays a role during the very final stages of predation, and allows the spider to see that there is an insect-sized shape in the middle of visual field, and that it moves.