Friday, December 12, 2014

Difficulty identifying Anyphaena dixiana

Usually when I come across a spider, I can tell which family of spiders it belongs to, but this spider baffled me. I found a female under a dog bed on my patio on November 23, 2014, in Austin Texas. Three days later I found a similar-looking male spider, and again I couldn't be sure of the family.

The female is on the left, the male on the right. These photos are not proportionally scaled: the female has a body length of 4.5mm and the male 4mm. We don't include the legs.

These two spiders have the general body shape and color pattern of a wolf spider (lycosidae), but they don't have the eyes of a wolf spider. Often a spider that looks like a wolf spider but isn't might be a funnel spider (agelenidae), but these eyes also were not a match for agelenidae. It is hard to see in these photos, but these spiders' eyes are in two rows of four, with all eyes about the same size. Funnel spiders have eyes in three rows, except for Tegenaria, which these definitely are not.

The only other spiders I could think of that had this general look and eyes like these are the spiders in the genus Syspira, belonging to the family miturgidae, and spiders in the family anyphaenidae. However, miturgids have heavy hairs along their tarsi and metatarsi (last two segments of the legs), while these spiders do not have that; and Spiders of North America: An Identification Guide (aka SONA) says that anyphaenids have first leg longest, while the first two pairs of legs in these spiders appear to be about equal length. SONA is the best compendium of spider taxonomy available for North American spiders. Other anyphaenids I'm used to have much longer first legs.

This left me unable to determine the family of this spider without looking at it under a microscope. I collected both spiders for identification but decided to give my friends on Facebook and Twitter a chance to guess the family before I scoped them for a definitive answer. Laura Lee Paxson said that it looked like the genus Anyphaena to her, in the family anyphaenidae. A quick glance at this genus on BugGuide did show quite a bit of similarity. However, no species listed looked like this spider.

Most araneomorphae spiders breathe through a tracheal spiracle that is located just in front of (anterior to) the spinnerets, in addition to breathing through a pair of book lungs, but the tracheal spiracle of spiders in the family anyphaenidae is located at about the middle of the abdomen. I usually take photos of the underside of the spider, again not using a microscope, in case there might be a useful distinguishing mark there. So let's have a look at the ventral abdomens of these spiders.

Notice that both spiders have transverse marks across the middle of the ventral abdomen. We won't be able to see the spiracle with these photos, but we might be able to see the fold of the abdomen in which the spiracle is found. There do appear to be folds parallel to these ventral marks, suggesting that these are indeed anyphaenids. I put both spiders under a microscope and confirmed that these are folds for the tracheal spiracle. This seems to confirm the family anyphaenidae.

But we're still left with a problem. SONA says that anyphaenids have leg 1 longest, contrary to these specimens. So now things get tricky. SONA is an attempt to organize a vast amount of information spread across hundreds of papers into a single reference that we can use to readily key spiders to family and genus; SONA is not authoritative. It's now time to look at the authoritative papers.

I decided to investigate Anyphaena. Norman Platnick wrote the most recent revision of this genus for North American spiders in his 1974 paper, "The spider family Anyphaenidae in America north of Mexico" (aka Platnick 1974). His description for the genus includes the words, "leg I not greatly elongated," suggesting that its leg lengths might not be noticeably different, contrary to most anyphaenidae. Anyphaenidae, and in particular, Anyphaena now seemed to be an option.

There are generally two ways to proceed once we know the family of the spider: (a) determine the genus before determining the species, or (b) determine the species and thus simultaneously establish the genus. Platnick 1974 helpfully includes the following in the description of this genus: "The genus is used here in a very broad sense; this prevents simple diagnosis, and makes detailed descriptions of each species group more meaningful than a description of the whole genus." This tells us to try approach (b). Under this approach, we attempt to find a matching species of Anyphaena before concluding that we have Anyphaena at all.

Species of spiders are most easily distinguished by their genitalia. Genitalia are often the most distinctive characteristic of species among both arachnids and insects. The majority of genus revisions are therefore keys to genitalia, and Platnick 1974 is no exception. Male genitalia are largely external and make for the easiest identification, so we start with the male spider. In males, we look at the last segment of the pedipalp, which we often lazily refer to as just the "palp." Here are two microscope photos of the male spider's left palp:

The photo on the left is the ventral view of the palp. The photo on the right is called the "retrolateral" view and is the side of the palp that faces away from the other palp. There are lots of parts. There were two drawings in Platnick 1974 that sort of looked like the left palp and two that sort of looked like the right palp. However, the paper includes the text, "Anyphaena dixiana [...] may be quickly recognized by the dorsal prong of the RTA" in males. This is referring to the dark tooth that you see at the lower right of the palp in the retrolateral view. The drawing in the paper shows a much smaller tooth, but I thought it reasonable to conclude Anyphaena dixiana because spiders exhibit variation within a species like all organisms do. Platnick 1974 also includes a decision key for identifying the spider from the genitalia, and that key also indicates Anyphaena dixiana. We have pretty strong confirmation for this species for the male spider and thus also the genus.

What about the female? I've only been guessing that they belong to the same species because they look similar and were found in the same location. The female genitalia is called an "epigynum" (normally pronounced "eh-PIJ-in-um"). It is located just behind and between the book lungs on the ventral abdomen. It is shiny in the first ventral photo above. Identification is sometimes done only by what can be seen externally, but it is often done by resecting the epigynum and looking at the hard ("sclerotized") structures on the dorsal side. We thus have a ventral view and a dorsal view of the epigynum. Here they are for this female, ventral first:

The ventral view did not look identical to any of the drawings in Platnick 1974, but there were two to which it was similar: Anyphaena catalina and Anyphaena dixiana. The paper refers to the dark rims on each side of the epigynum as "sidepieces." The sidepieces were much farther apart at the bottom (posteriorly) in dixiana and much closer together in catalina. This specimen's sidepieces were halfway between those of the two drawings, leaving us without a species identity. However, being halfway between these drawings clearly placed the spider in the genus Anyphaena.

The paper also provides a dichotomous key -- a decision key -- for helping to identify specimens. The key emphasizes the "hood" of the epigynum as well as the "midpiece." The dark sliver at the top of the epigynum is the "hood," and the subtle bump at the center is the "midpiece." The relative width of the hood to the sidepieces narrowed the epigynum down to eight different species, including Anyphaena dixiana. The key further distinguishes between sidepieces that are "widely separated posteriorly" and those that are "approximate posteriorly," between sidepieces that are "very wide" and "narrow," and between midpieces that are wider than the hood and those that are not, among other things. The drawings are intended to be used as a guide for gauging relative sizing, but we're using the key because relative sizing isn't distinctive in this case. The key appears to narrow us down to one of the following: Anyphaena dixiana, A. gibboides, A. autumna, or A. judicata. Checking the drawings again for these species, we find that our epigynum looks nothing like any of these except Anyphaena dixiana; the female therefore keys out to Anyphaena dixiana.

This does not yet mean that the female is Anyphaena dixiana. Because none of the drawings seem like a perfect match, it was also possible that I was looking at a new species, not previously described. That can happen; the decision key only works for species that were known at the time it was written. I had the option at this point to just assume she was Anyphaena dixiana and move on, leaving it to a specialist later to double check. I also had the option of confirming the identity by removing the epigynum and looking at the dorsal aspect. I like to be cautious and opted for this second approach. The second epigynum photo above shows the dorsal side. This view appeared similar to the drawings for three different species: Anyphaena dixianaA. gibboides, and A. catalina. The ventral aspects don't match for the latter two, once again suggesting A. dixiana and helping to confirm our educated guess that was based only on the ventral aspect.

It remains possible that the female is not A. dixiana, but because she was found near a clear male of the species, odds are that this is the correct ID and that we are witnessing the genetic variability of the species. I decided to also call the female Anyphaena dixiana.

And so you see some of the challenges of identifying spiders, all the way from family down to species. It helps to have a good sense of as many genera as possible to help rapidly narrow the possibilities, as we saw with Laura Lee Paxson's educated guess about Anyphaena. Sometimes we have to work backward from genitalia to even be sure we know the genus. And sometimes we are not going to have a solid match and have to make a call about whether what we're seeing could be within the genetic variability of a species, leaving it to a specialist later to make a more confident call.

One final thing. During the photoshoot, the male kept rolling into a ball when disturbed and staying that way. I could blow him around, and he'd roll across my bathtub studio. Poor fellow.

You can find more photos of these two specimens on BugGuide: the male, the female.

Saturday, December 6, 2014

Problem-Solving Strategies

Here is a list of problem-solving strategies. Most of these strategies arose from reflection on how I develop software design specifications. This is an evolving document.

  1. Identify and attack tangentially related problems. This often sheds light on or even resolves the original problem, even if the process happens to reveal more problems.
  2. Create solutions gradually by iterative refinement. The best solutions are evolved. Solve select problems, leaving others unsolved. Gradually throw more solutions into the mix. Don't expect solutions to grow by accrual, as periodic complete transformations may be necessary. Each new iteration still benefits by being derived from preceding solutions.
  3. Find and depict specific examples of the problem. Articulate the problem separately for each example to find a common articulation or to learn the problem's component pieces.
  4. Find the right questions to ask -- about the nature of the problem and what is desired in the solution. Asking questions about the smallest things that you don't understand can open up whole new spaces.
  5. Have sessions of diving in intensely. Think, organize, and re-think, over and over, and don't quit until you find yourself having trouble focusing. The objective is to tour some of the space and not to leave until you get some sense of it on which to chew, even if it's a sense of its chaos. These sessions complement more scatter-brained sessions for breaking road-blocks.
  6. Draw pictures at every opportunity. The act of trying to find a picture that captures a problem or solution is often enlightening. Watch for what you know and what you don't know.
  7. Apply aspects of possible solutions in the real world and see what you learn. Create prototypes and test them out.
  8. Talk out loud to yourself at every opportunity. Hearing it spoken engages the brain as if it's listening to another person.
  9. Play with speaking fast and unfocused. Just start talking, out loud, and don't let yourself stop. See where you go. This can help bubble up ideas that are just below the conscious that you may have been having trouble grasping.
  10. Talk with other people who can help and spend time considering their input. Sometimes just trying to explain something to someone can bring clarity.
  11. Attempt to explain both problems and solutions in writing, with precision.  Areas where you have trouble being precise may be areas where you lack understanding, where decisions may need to be made and problems solved.
  12. Don't be afraid to walk away from problems for days, weeks, or even months. Give yourself permission to temporarily **quit** the problem, with knowledge that you'll later be returning to it. If a better problem statement or even a solution doesn't pop into your mind during the break, you're likely to have a fresh perspective and more enthusiasm at the end of the break.
  13. Fearlessly start documentation and product over and over again instead of trying to backfit new understanding and new language into old expression.
  14. Label aspects of both the problem and the solution with the most precise words possible with no compromises. Words affect what you think about and how you think about it.
  15. In your own notes, express ideas in the most succinct language available to you that requires minimal or no effort on your part to decipher.
  16. Study/scan related information, domains, or technologies for ideas.
  17. Look for apparently similar problems in unrelated domains and explore mapping their solutions back to your domain. Ask yourself, "What does this remind me of?" and be open about your responses.
  18. Experiment with looking for answers in random things you see, allowing yourself to think crazy things. Sometimes a new property of a potential solution becomes available.
  19. Write long lists of ideas. Sort them, analyze them, and then write new lists.
  20. Articulate your *feelings* for the solutions on the table. This helps with identifying personal biases that interfere with solving the problem but have nothing to do with the solution.
  21. Periodically restate the problem throughout the day, as you're doing other activities. Just restate the problem, allowing yourself to find new words each time, but not requiring that you do so. State the problem as you lie in bed falling asleep. Allow your mind to wander around the problem when you wake in the morning, before getting out of bed. Sometimes that's when the answer comes. Standing may dispell the thinking.
  22. Play, "What would so-and-so do?"  Pretend that you are another person and wildly allow yourself to spit out answers from their perspective.
  23. When you've had enough planning, brainstorming, and problem-solving, say "Screw it!" and just try something to see what you learn. Play, play, and play. Putz around some too.
  24. Listen to stimulating music, but only instrumentals or lyrics in languages you don't understand. Or take a walk through through the woods where you won't bump into people who might distract you.  Or dance around and around as you think. Walk about, don't just sit there.
  25. Eat lots and lots of chocolate. (Or perhaps coffee is your thing.)
  26. Know that all problems fall with the right leverage, sufficient pressure, and enough time.
  27. If you sense a slight discomfort with some aspect of the solution, try to precisely articulate that discomfort. Doing so may clarify your understanding of the problem or the solution, or it may identify a problem with the solution.
  28. Analyze the problem from every angle you can until you find yourself stuck. Then start solving the parts of the problem that you do understand. You're likely to improve your understanding of the difficult parts of the problem in the process.
  29. Solve the problem for a specific case, even if oversimplified and unrealistic, even if it leaves other known problems completely unaddressed. Then generalize the solution to handle another case. Keep generalizing until you've handled all the cases. It's helpful to first enumerate the cases that need to be handled.
  30. Work to discover the rules of the system. Keep track of ideas on possible solutions as you go. Once you think you have the rules, line them up to see all at once. Generalize your space of ideas as broadly as possible, and then sequentially whittle down the space for each rule.
  31. Try again. You haven't failed until you quit trying.

Thursday, December 4, 2014

The True King of the Sky

The robber fly is usually considered to be the top aerial predator of the bug world. These photos may dethrone the robber fly and establish the spider as true king of the sky.

This past summer, I visited an Argiope aurantia (aka "Black and Yellow Garden Spider") in my front yard day after day to see what she was up to. She had made her web below some cables that string to my house. I often saw a robber fly monitor the sky from the bottom cable. I thought it was a cool sight and took this photo on August 7th, 2014.

Three days later, on August 10th, I found that my spider had caught a robber fly. This isn't a great photo, but it shows the spider's web near the cables on which the robber fly would perch. She is feeding on a robber fly in this photo.

This close up makes it clear that her prey is a robber fly. Out of curiosity, I kept looking for a robber fly on the cable for days afterward but never saw one there again. I suspect that my first photo is of the very robber fly that she is devouring.

So is it rare that an Argiope would catch a robber fly? Perhaps not. The following photo is of another robber fly in this same spider's web, taken on July 19th, weeks before I took the above photos.

The robber fly may be the top insect predator of the sky, but spiders are the top arthropod predator of bugs that fly. Maybe spiders shouldn't count because they themselves don't fly? But spiders do fly! Mostly young spiders fly—by ballooning—but some adults have been seen ballooning too.