the daily discipline of emergence checking

So, I’ve previously described the tube experiment, which I’ve been running for about a month now.  Each tube is filled with detritus and water from the leaf where it sits, and contains larval insects. The larvae are in pairs of (congeneric!) species wherein one is a generalist on all bromeliad sizes, the other found only in the largest plants.  If there is anything in the environment provided by large bromeliads that makes the specialist succeed over the generalist, then this experiment should pick up that difference.

The larvae in the tubes eat the detritus, experience the environment, and, if they do well there, eventually pupate and emerge as adults. When necessary, there’s a little popsicle stick there for them to crawl up to freedom.

Of course freedom, in this case, is a net bag where they wait for me to collect them with my aspirator!  The taxa of these bromeliad animals are poorly known, and there was just no way to identify the adults positively when they were still alive.  So, I’ve been catching every adult chironomid that emerges, killing them, and preserving them in eppendorfs atop some silica and cotton.  So far I have well over 150!

It’s strange the skills you acquire as a field biologist.  I spend at least an hour a day at this task, and over the last month I’ve gotten pretty good at deftly slipping the mesh top off and seizing the insect with my aspirator.  I’ve also become embarrassingly attached to this particular tool of the trade, which I made myself from a description by Karen Needham, the entomology curator at the Beaty Biodiversity Museum.

The routine of checking the tubes is a classic field-biologist repetitive task.  Most of the tubes are empty, but each has to be checked every day, and maintained carefully (did you know that ants love to eat mesh and elastics?  I was surprised, too).  Emergences are exciting — they mean data, and are especially essential in the two-species tubes which test for competitive differences.  Emergences are also a trial, and mean that the collecting trip will be prolonged while I catch the insect. On big days (once there was 16 in a single day!) the whole collecting can take close to two hours, and my field assistant would let out a shout of “NOSSA!” (the Brazilian equivalent of “jeez!”) at every fresh discovery of an insect.

Someday soon I mean to sit down with the adults and try to work out if the guesses I’ve been making all along about species identifications of adults are actually accurate. Fortunately I have lots of monoculture insects to learn from.  I’ll study them, and then see how I can do with the polyculture insects!  We’ll see, and I’ll let you know what I find!

There was a naked woman on the UBC campus

The young lady in question took off her clothes in peaceful protest of an anti-abortion group protesting on campus.  You can read her description of the events on her blog.  I won’t rehash the information that you find there, other than the barest details.  The anti-abortion group was using graphic images of genocide as a shock-tactic to highlight their views on abortion.  Ms Davidson removed her clothes in protest, as she says, because “My body is where I exercise and appreciate my freedom on a daily basis, and I reject outright the assertion that by supporting the right to free, safe abortions, I am turning it into a tool of mass murder.”

The university, unfortunately, has responded negatively to her action, and is charging her with ‘student misconduct’.  They seem particularly incensed that she ‘disrupted’ the display of the abortion group, while (according to her side of the story) she did no such thing, and the group has not even complained.

She urges all those concerned by the University’s response to email the president, Stephen Toope (, and Chad Hyson (  I have done so (email below) and encourage you to, as well!


President Toope,

I am writing to express my concern over the reaction to Justine Davidson’s peaceful protest on Thursday, 08 March.  Her excellent and well-written description of the events that provoked her protest, and the intelligent thought that went into her action, are a great credit to the university.

Such peaceful protests are a sign of our University’s status as a place of discourse, and their existence makes me proud to be a graduate student at UBC.  I urge you to reconsider the allegation of ‘student misconduct’ directed at Ms. Davidson.  Her protest, as far as I can tell from the version of events on her blog, has not disrupted the protest of the GAP group which was demonstrating at the same time.

I was not on campus to witness these events, and I grant that my information is derived from one source only.  However, my first reaction on hearing of Ms. Davidson’s action was one of pride in my University for being an institution where such acts can occur.  A member of our community has taken courageous action motivated by her ideas and convictions, which no doubt were partly formed at UBC.  This is precisely the UBC I wish to attend.  As our President, I urge you to show a similar courage in celebrating, not condemning or punishing, her action.

Best regards,
Andrew MacDonald
PhD Student
UBC Zoology


Meet the Tube Experiment!

I’d like to introduce you to one of my experiments.  This is its first time out in public, so here goes!

The background for the experiment is a simple story.  The bromeliads vary greatly in size, from tiny plants whose whole contents would fit into a coffee cup, to massive ones with a capacity of well over two litres. Two years ago, my supervisor dissected over two dozen bromeliads from this habitat, covering the full range of variation in size.  When I looked at this data, I saw that some animals were only found in very large plants, while other animals have a much broader palate.  One way to look at this is to predict the probability of finding a certain species as a function of size.  The resulting graphs look a bit like this:

I had fun, for a while, drawing lots and lots of these curves in R.  I thought it was particularly strange that sometimes very similar species were so different in their preferences.  A particularly striking difference occurs between the midge species.  Midges (Chironomidae), have aquatic larvae, while the adults are mosquito-sized flies.  They are sometimes used in environmental monitoring — but I’ve also heard them ridiculed in talks as species which are pointless to distinguish, which could not possibly differ ecologically.  They’re so similiar, the argument went.  How could these little red worms actually be that different?

However, time and again we’ve found the same pattern that my supervisor did two years ago.  After a month of observing these animals often, I can identify them fairly quickly.  I was identifying them for a friend, and I turned around and said “This bromeliad was really large, right?  and this other one was tiny?”  He replied that yes, the first was over 900ml and then second less than 150 or so.  I indulged myself in feeling like an Ecological Sherlock Holmes.

So, this brings me to the experiment.  Reciprocal transplants along a gradient are staples of ecology, so I decided to do one myself.  The idea is, you collect the larvae, and you put them in these convenient plastic tubes.  The tubes have holes in the side to let water through (though covered with mesh to keep unwanted larvae out and experimental larvae in).  Then, you put the tube in the bromeliad.  They are thankfully roughly the same size as a bromeliad leaf-well, and sitting there they receive all the same environmental conditions as the rest of the plant.

I’m doing this with two species: Species One, a picky, large-bromeliad-loving kind of chironomid, and Species Two, the easygoing, generalist counterpart.  The treatments are a simple checklist of all the usual suspects when it comes to determining where a species occurs and where it doesn’t:
Is it the environment?  Let’s put each species on its own, in large and small plants, and see how they do.  The large bromeliad species should do poorly in the small habitats.

How about competition with each other?  Maybe Species Two is found in small bromelaids as a refuge from fierce competition with Species One. Let’s put them together, and see who wins!

Maybe predation?  Predators as a group are only found in big plants.  Maybe the small-bromeliad species can’t handle the predators, while the other one can?

These treatments get replicated in large and small plants.  Every day I check what emerges, and try to catch them.  This is a laborious task, full of challenging catches and disappointing losses, but I will describe all that later!

randomize or block?

Let me pose a hypothetical situation to you.

(Any similarity  between the experiment described below and any real experiments, researchers or analyses are probably mostly coincidence)

You are designing an experiment.  Your experimental organisms, let’s call them Mathom smialis, have been ordered for the purpose.  You have four treatments, and you’re planning on setting up five replicates of the experiment, so you’ve ordered 20 Mathom.

The trouble is, they vary somewhat in size — not as much as they would in wild populations of Mathom, but enough to be noticable.  You’re afraid that your response variable — let’s say, the diversity of Stoors which colonize the Mathoms — may be an increasing function of size.  So, of course, you measure all your Mathom before you begin.

But what do you do with that information?  You could assign your Mathom to experiments randomly.  Write down your 20 replicate x treatment combinations, scramble them up, and assign your individuals to each.  Or, you could try to keep the variation in each replicate the same: do the four treatments once with very small individuals, then again with larger ones, etc.

What worries me about the second option is that it’s inviting block x treatment interactions.  What if in the block with larger individuals, treatment differences become really large?  If larger Mathom  can collect a greater diversity of Stoors over their lives, then treatments that increase their diversity have a larger scope for effect in replicates with larger individuals. In other words, treatment A could cause an increase of 20% among small individuals, but a 60% increase in larger ones.  This makes the blocks not true replicates of each other.

I think the first option (simply randomizing) might not appeal to researchers who want to minimize variation in their response.  But minimizing varation within blocks is not as important as creating blocks that are true replicates!  So I’d come down strongly on the side of completely randomizing.  Does anybody else have any other thoughts?


Cardoso with colleagues!

More scientists have arrived!  Two of my Brazilian colleagues are here, each with a field assistant, making us 6 in the house!  I’m enjoying talking about the finer points of bromeliad ecology with these guys.  Not to mention the extreme Portuguese practice I’m getting!  But most of all, I’m enjoying having other people who are obsessed with the system, too.  When I describe my work to friends back in Canada, I’m obliged to put it in general terms, or risk some very bored looks.  Here, I can indulge in some get excited about our favourite method for collecting scirtids, or the traits I use to separate two similar morphospecies, without fear.

I did the bare minimum of actual fieldwork today: emergence checking only.  Unfortunately nothing emerged at all!  Yesterday’s crop of adults was very low, too. I’m beginning to think the experiment is over.  If so, emergence has been low: 120 adults have emerged out of over 600 which I placed in the field.

OK, before I to to bed, I just wanted to share today’s fun Natural History moments:

We came home to fine the backyard boiling with army ants!  This was interesting, until they started invading the house and had to be repelled with brooms and insecticide.

In the Restinga today: three large Rusty-margined Guam, *Penelope superciliaris*

At night: a Massive and Interesting Beetle (Scarabaeidae), which was trying to get in with such force that it shook the screen door.  We through it outside, and were treated to the spectacle of a Burrowing Owl pouncing eagerly on it, bashing it about a bit and flying off.

One final thing.  This is mostly a note to myself, but I wanted to include a recipie for farofa, in case I forget it:

fry bacon in butter, until browned.
add more butter, finely chopped onions.  fry till golden.
season with oregano, basil and hot sauce.
add coarsly-chopped plantains and stirfry till the plantains soften slightly, turn color and start to smell nice.  DO NOT let them break down.
Add manioc flour in large scoops.  Gently fold in, trying not to mash the bananas.  stir continuously until the flour is lightly toasted.  add enough so that it is rather dry.
Eat with roast beef, beans, or anything with yummy sauces that need soaking up.


survey results!

Guess what: it didn’t rain last night!  I tossed a laundry basket onto
the sidewalk outside, and this morning it was completely dry.  So, I
downed some coffee and went out to start measuring.  Under an overcast
sky, I pushed a wheelbarrow of equipment out to our “Field Lab”
(actually the bathrooms behind an unused cafeteria.  But it’s the only
available faucet, so what can we do?).

I found a patch of bromeliads that hadn’t been molested in the past
(meaning dug up and replanted or washed to collect insects or discarded from
an experiment, or any of the other indignities we ecologists visit on
these little plants).  I decided I would choose bromeliads all from
the same location, so that the only difference would be their
different sizes.  The same canopy, the same amount of sun  — the
plants may even have been related, since this species spreads by

Of course I didn’t do as many as I imagined.  I marched out thinking
“I’ll try to get through 30 before lunch!”.  How did I figure that was
possible? To measure a bromeliad’s volume, you have to go the the
field, carry the plant back, dump it out and carefully collect that
water.  Measuring this volume gives you the “actual volume”, meaning
how much was left in the plant after evaporation. Then, you pour a
known volume of water into the plant, collecting what runs out into a
bucket.  You carefully measure the volume of this runoff, and by
subtraction calculate the plant’s volume!  Finally, you replant the
bromeliad, collecting another one as you do so.  The whole process can
take 20min for a small plant or 30-40 for a larger one. 


I did, however, manage to get through ten plants!  I returned home in
time for lunch — but I had to stop anyway.  All morning the weather
had been threatening rain — but halfway through my first bromeliad I
looked up and saw the island’s mountains, “the Three Brothers”,
completely hidden by a curtain of rain! I managed to get the plant
done and rushed home as the storm finally, finally broke over me.

As excited as I was to do this work today, we did very sorely need
rain!  The bromeliads were so dry yesterday that it was impossible to
insert any instruments to measure oxygen or pH, and in my tube
experiment I was getting hardly any emergences at all.  But today in
the afternoon, I walked back to the restinga (this time with a
raincoat) and found several adults had emerged!  I have a hunch that
they only emerge during or after rainy days.  I’ve started keeping
more careful notes of the weather so I can match it up with my
emergence data.

This reminds me that I ought to describe my other experiments!  I
should devote a later post to each of them.  But today I wanted to
show you the results of my quick little survey:

There is my morning’s work, in two flavours (log-log and
vanilla). Looks pretty linear, eh?  To solid line is the 1:1 line
(i.e., completely full bromeliads); you can see just how little water
the bromeliads contained, relative to what they could hold!  The black
dashed line is a linear fit to the data, and the red dashed line is a
smoothed fit. 


So apparently, bromeliads dry at a constant rate no matter how big
they are!   What do you guys think? I was wondering if this is really enough data, or if the analysis is quite right. I’d love to hear your thoughts!

Data Day

So, today was a day very much, once again, devoted to data-entry.  I went through all the datasets that my (very organized) assistant is helping me manage.  The morning spent on the process. In particular, to fighting with the format of the dates.  It’s frustrating and satisfying at the same time.  In the past, I’ve sometimes not entered data until long after it was collected — always with much self-recrimination.  This time, I’m keeping the advice (from a biostats book) in the front of my mind: “Enter your data as soon as possible”.  It is also much more efficient to set up good datasets early: then your assistant can help you enter data!

Part of the issue is that immediately after you collect data you remember all the little details: the sample that was partly dropped, the funny larvae that you’re only 70% sure fits in this morphospecies.  Later, the quick notes in a field notebook don’t convey all that information.  I spent some time having great fun following Ben Bolker’s advice and adding comments to my .csv files.  (Readers of this blog will quickly learn that I am an obligate R user, always keen to learn another trick!).  You can put comments in anywhere in a dataframe: above or below your data, even in the same cells.  Then, you just use read.csv() like normal, but with comment.char=”#”.  I’m going to use this especially to write down what my column headings mean, what units they’re in, etc., at the start of a dataset. I’ll toss in a screenshot to show off, I mean demonstrate, what I mean.  It’s full of comments, but it reads in like a perfectly normal dataframe!


OK enough computer talk!  Today I had an idea that I wanted to share with you.  It is perhaps a doomed idea, so at least I’ll get to talk about it and it will have some kind of life.  So, it has been **extremely** dry here.  Which I admit is nice for working and for drying the laundry — but after all, we do work in an aquatic system!  Some water would be nice!  But it occurred to me that this is an opportunity.  I’m curious about the effect of bromeliad size on the community they contain, and I’ve noticed that the small bromeliads have almost no water in them anymore.  Do small plants dry faster, or do they simply have less water to dry?  Of course it’s possible to measure that with an experiment, but a quick and easy way would be to measure the volume of lots of bromeliads with different volumes *at the same time*. I’d measure how much


water they have in them already (by dumping them out and measuring the water in a cylinder) and then fill them up to see how much they CAN hold.  of course all the points will be below the 1:1 line, since bromelaids are usually partly evaporated.  But the shape of the relationship below that line should give us a clue to the question “do small bromeliads dry faster or slower than big ones”.

I say this idea is doomed, though, because it would work best after a long period of drought, when lots of bromeliads have lost water.  However I’m sitting here, late at night, and listening to distant thunder.  Will this minor idea be doomed before it has a chance?  Even if the weather stays dry, will more pressing projects prevent it from happening?  Stay tuned!

Hot Science!

Today was a day of hot science.  And not simply because it was 30C in the shade.  No, because today we finally got serious about an important part of the project here: measuring the sensitivity of the insects to heat.  I’m going to try to upload a photo to the blog, for the first time, so that you can see what our apparatus looks like.

The equipment consists of a hot water bath, tricked out with duct tape, wire, a thermometer and, yes, a clothespin.  The clothespin keeps the thermometer in the water at the correct angle to make sure its immersed properly.  We need the thermometer, because this particular water bath doesn’t tell you the temperature to which it heats: the dial simply has the numbers 1 to 6. 

The wire grid cunningly affixed to the top of the water bath holds the tubes for the insects.  There are ten tubes in all, each made of plastic and each containing a little insect.  Holes in the side of the tube allow the water to flow between the tube and the hot bath.  I know this seems ideal, but I confess candidly that these tubes were not designed for *exactly* this purpose but for use in bromeliads.  Still, they are perfectly suitable for this.  The word “jury-rigged” comes to mind.

Anyway, after the apparatus is set up the dial is slowly turned up, and the water heats.  When I say slow I mean SLOW: my assistant turns the dial a fraction, just a few degrees of the circle.  And then he waits with a stopwatch, watching the temperature climb and checking the insects for ‘loss of muscular control’. After three minutes, he advances the dial another fraction and waits again.


How in the world can you tell if a damselfly larvae has a “loss of muscular control”?  And other profound questions.  Well, it turns out that it is fairly easy to define.  I heard a physiologist friend use the term “ecologically dead” to refer to a state that might not in itself be immediately fatal, but that would probably get you killed quickly in nature (Think of fish floating belly up, not dead but not at all comfortable).  Our damselflies actually do go belly-up, legs twitching erratically, losing their grip on the wall of the tube. 

PLEASE NOTE: all insects make a complete recovery and are returned to bromeliads after their ordeal!

I just realized that I forgot to explain *why* two eager young scientists are sitting on a tropical island, intently simmering larvae to near-death.  It is because we are trying to
a mystery: why are bromeliads in the open habitats so different from those in the closed?  For example, we have two damselfly species, very similar in many ways.  However we find both species coexisting in the closed habitat – but only one in the open!  Is it because the other species is too sensitive to the extreme temperatures that the hot, sunny, open habitat receives?

So far results are only a little encouraging.  While we now know exactly how much hotter the open restinga gets — cool graphs coming up! — the critical temperatures of both species are very close together, with our suspected “more sensitive” animal being only slightly lower.  The current plan is to increase the wait time between temperature increases, to see if that helps to get more resolution.  And to test more species!


This morning we washed a record number of bromelaids.  This is a messy process: we go to the forest, take two large bromelaids, and shake them upside-down into a large plastic tank.  then we scour the bromelaids with a hose, until there is no more detritus in each of them.  Then comes the rather strenuous process of sieving, where the water from the tank must be poured into a smaller tub, and then poured through a sieve.  the bromeliads are replanted, and the water that has fallen through the seives is poured into the plants.  We have two seives; today we kept only the ones which stayed on the coarser sieve, since the insects we are seeking are rather large. We needed 12 individuals, so we got up early and washed bromeliads all morning.  We joked that we would have many times what we needed.

We found only one!

The animals we are seeking are two species of beetle larvae in the family Scirtidae. There are two similar species here (in fact comparing them is the very subject of the experiment), but they are so similar that they can only be told apart in the later stages.  We caught many individuals that were small and ambiguous, some which were old enough to be clearly “Species A”, but hardly any that were “Species B”.

My hardworking and sincere field assistant was kind of downcast about the whole thing. Personally, I’m eager to get to my other projects here.  I’m OK with putting this experiment to rest with the five replicates that are already in the field.

Anyways, tomorrow is a day off!  We’re going to Maruja, a small fishing village on the other side of the island.  The next post will be all about our adventure!

Salad Day

Salad Day

Today was field-work salad: a mix of all the activities that dominate daily life here. The metaphor is easy since today was the day we finished the last of our precious lettuce, which we bought in town yesterday.  So, no more nice crunchy lettuce for another week!

Today was a day of reorganization of our lab space.  I’m beginning to accept that if you set up a field lab for any considerable period of time, you have to set aside a day every now and again to move things around, to adapt the workspace to your working habits.

Today was also time for Larvae Husbandry.  Some of the larvae live in special blue-capped plastic tubes, where we hope they will soon become adults.  We protect and feed them and cheer them on daily.  Just recently we got a syrphid fly adult, and we were overcome with rejoicing.  I am looking forward to finding even more adults, and eventually overcoming the taxonomic impediment that we have here. 

I also checked my experiment for adult insects.  Both today and yesterday were sunny and very, very hot; I’m beginning to wonder if less insects emerge in such days.  Between yesterday and today I only caught a handful of adults!

Like a salad, this post is going to have to serve as an appetizer: I’d love to tell you more details, but I’ll need to devote a post to each activity we are doing here.  But I need to get to sleep!  Because if today was salad, tomorrow will be steak: tough, bloody and satisfying.  We’re going to wash 10 bromeliads, collect insects and set up one (two!?) more replicates. 
I’ll let you know how it goes!