Skip to content

Interesting papers 11

Communities in cold regions enhance warming effects on soil respiration (Karhu et al 2014).

Road building strategies key to sustainable development? (Laurance et al 2014)

Bacteria avoid growing until antibiotic concentrations are low (Fridman et al 2014).

Toenail dwelling microbes similar to those found in moist skin areas (Oh et al 2014).

You are responsible for the microbes in your home (who/what else would be???) (Lax et al 2014).

Applied evolution can save the world (Carroll et al 2014).

Greater stability of floral resources in more diverse communities (Dorado & Vázquez 2014).

Density dependence irrelevant for good predictions of population dynamics (Wootton and Bell 2014).

Plastic populations evolve more (Schaum & Collins 2014)!

I’m a post-modernist pragmatist. You? (Moon & Blackman 2014)

The bad side of corridors (Haddad et al 2014).

We need more individual level trait data (Kraft et al 2014), like this (Kremer et al 2014).

Easy size spectrum models, in R (Scott et al 2014).

My first flipped classroom course

Flipped classroom courses involve students watching lectures in their own time, and working on assigned problems in class (more info. in this Wikipedia article). Summer 2014, I and three colleagues decided to develop such a course to teach R and some statistics at a basic level. We were each already teaching our own traditional format courses, so thought to both rationalise this teaching by sharing our resources, and to do so via online learning, which led us to the flipped classroom. We were also influenced by the rise of MOOCs and their online learning methods, but aimed for a course that was for UZH students (initially at least). Given those aims, the following questions and answers arose.

What structure for the course? We developed a structure with two types of online lesson: Problem Studies and Skills Lessons. The Problem Studies present and solve a real problem, and also communicate the solution. They do not go into details about R or statistics: this is the realm of the Skills Lessons. For example, we have a Problem Study about the relationship between income and orgasm frequency based on real data, and a Skills Lesson about Preparing Data. This course structure attempts to keep students focused on understanding and solving problems, and efficiently communicating the answer. Proficiently using R and statistics is necessary but not the focus, so it goes in Skills Lessons, so as not to distract during the Problem Studies. So far this structure is working well. Aside from that, the edX advice about lesson structure was very valuable (e.g., each lesson consists of multiple short [5-10 minute videos] with a few somewhat formative questions after each).

What online learning platform to use? We chose Open edX for at least two reasons:

  • It is proven, both for instructors and students. It works and is pretty easy to use.
  • Creating courses is easy, and there’s good documentation and support. There’s even a course about creating edX courses.
  • It can “freely” be installed on a University server, which provides the potential for complete control of learning material and student information. (Freely in quotes, since although the software is free, someone has to install and maintain it.)

How to do the videos / screencasts efficiently? Preparation is key. I found an hour writing a good script meant I could make the video / screencast in one or two takes. When I didn’t prepare a good script, it was a mess that either required lots of video editing, or couldn’t be fixed at all. Bit of advice… dig out your personal photos and videos, and make a home movie. Doing this will either: give you get nice home video and knowledge of how to use video editing software that can be used to prepare you flipped course material; or a crappy home video (even no home video) and knowledge that you should prepare really well, so you need very little editing.

How to make screencasts? I can only answer from the perspective of a Mac user (though was not so difficult for my alternatively blessed colleagues). As such I use Quicktime Player to record the screen and audio, and sometimes Photo Booth to simultaneously put may face in the screen, so students can see me at the same time as whatever else is on screen (only problem is I can’t figure how to resize the Photo Booth window, or get rid of everything apart from the camera image). I use a nice microphone (e.g., Apogee Mic 96k) and a mic stand (rather than stand it on my desk) to isolate the mic from the sounds of me typing, clicking, and doing other things on my desk. I try to record in one or few takes, not being too fussy about little mistakes. This cuts down on editing time, as well as being quite adequate (no complaints so far). I then use iMovie to do any little edits, to make small adjustments to sound levels, and to make an mp4 which I upload to Youtube. Be sure to make an at least HD mp4. All this I do in my office when I know it will be quiet, and I switch off all electrical devices I can (e.g., external drive, printer, silent phone, Skype off).

UPDATE: Looks like Camtasia might be very good for making screencasts, in particular as it can capture the view of the HD camera at the same time as the screen.

How to make the videos of yourself in front of a black / white board? I use my iPhone, the same mic as above, with an appropriate stand for both. Be careful that the iPhone focuses properly on you and the board. Make sure your writing is big enough. Try to do one take, so you need less video editing. Test the sound levels.

Was I smug about using a Mac? Yes, I was. But I tried to hide it :) And my Windows and Linux colleagues found good solutions also. So probably I had no good reason to be smug.

Where to host videos / screen casts? Somewhere reliable and able to handle however many students you expect to have, all watching at the same time. Before you invest lots in a solution, check the url of the video works when pasted into an edX lesson. We are currently using YouTube (which means we don’t have total control of course material), as its easy and works.

How did we avoid lots of tech problems? We each made about five of test videos and screen casts, put them on edX, and looked at the result. Thus we got to see and test the whole process, and troubleshoot it, rather than making 20 lessons only to find something was wrong with them. Also useful, was making one lesson (i.e., getting it fully working on edX) at a time. Again, this avoided any large scale screw ups.

What did I do wrong? I wore a variety of shirts in different videos. I ended up with a lesson where I magically grow a beard, changed shirt, and lose hair during a short break. I recorded several videos out of focus, some with sound levels too low, I messed with sound too much in iMovie sometimes, so had to export the mp4s again. I tried to get rid of background noise after the recording is done, e.g., via the background noise reduction in iMovie. It takes all the life out of a voice. So record where you have little background noise, and or use a lapel mic.

How bad is it that the four instructors use three different operating systems (1 mac, 2 linux, 1 windows), and different methods of interacting with R (e.g., Rstudio, base R, emacs speaks statistics)? Not too bad… the students haven’t given much negative feedback, though some has suggested they’d prefer everyone using Rstudio — in which case OS is less relevant).

Should we make the course a MOOC? Not at the moment. For one thing, the time in class, chatting face to face, seems really important. I recently heard about an Intro to R course with ~37’000 students enrolled. I’m not sure we’re ready for that!

Will it become a MOOC? Probably not. We are, however, thinking of giving the material / course away, but haven’t yet got it to a standard at which we would be happy to do so. There are also issues with this that we haven’t thought hard enough about.

Did it work? Yes, in the sense that the students seemed to really like watching lectures online, as they can pause them, can watch them again, can watch them at 1.5x speed (or half, or double). The in-class session were fun for me, and it seems the students. Most had done the work required, and engaged very well.

Was it all worth it? Yes, and it wasn’t so difficult anyway. Fun to make the course, especially collaboration to do so, nice to try and learn something new. Great to be finally comfortable listening to and watching myself (I sound and look OK!).

Does it save time? The point isn’t to save time… its to give a different and hopefully better learning experience. But to answer the question, I expect its about the same overall time investment, over 3-5 years, as preparing and running a standard format course.


Some notes on technical issues, after a second round of recording.

I got an iMac with 5K screen. Seems that using quicktime cannot now make a recording with well synced photobooth window. So I recorded the screen with quicktime, and my face with iPhone. Worked mostly fine, though extra work in iMovie, to put video of face within the screencast, and to sync them.

I was recording in summer in a warm office, and with the recording work, my computer fan came on. Some background noise on the video. Though iMovie can get rid of this, it takes some life out of vocals.

Finally, have some trouble with always getting 1080p HD. Sometimes iMovie will only output in max 720p. Will update when I figure out what’s going on.

Checklist in iMovie:

  • Trim and sync screencast and face videos.
  • Position and flip face video as required.
  • Audio off on face video, audio auto on screencast, consider background noise reduction.
  • Export at 1080p if possible.

Interesting papers 10

Importance of rapid prey evolution (Hiltunen 2014).

Complementarity increases herbivore control and redundancy stabilises (Peralta et al 2014).

Multi-host multi-parasite networks are a research priority (Lively et al 2014).

Warming can result in enemy release (Fey & Herren 2014).

Indirect effects of climate dominate, again (Michaletz et al 2014).

When should I start giving?

Give and Take, Why helping others drives our success, by Adam Grant… what a great read. It explains that givers are often the most successful people around, and why. It can also be interpreted as a manifesto for change: I already helped people that I might not have before reading it.

But when should one start being a giver, or give giving more time? Is it better to be a matcher (or taker) early in one’s career, then gradually put more resources into giving? Clearly one must perform well early on, and recognise and tend to important short term goals (publishing your first few papers, writing your thesis). On the other hand, I suspect important short term goals will always exist. And perhaps more importantly, giving is probably akin to investing in a pension. Early investments are more important than later ones, due to the way interest builds on interest. If the benefits of giving have properties of geometric growth, its never to early to start.


start early

Graph: Two exponential growth curves with the same rate of growth, but one starting a little later than the other. The effects of this delay just grow and grow.

That said, I’d never advise a PhD student (or postdoc) to compromise publishing a paper. But also I’d be surprised if there wasn’t time for a little smart giving, to get that investment started.

Ecological Epigenetics PhD


Biodiversity faces many threats. Predicting the consequences of these for biodiversity requires an understanding of effects across multiple levels of ecological organisation: genes, individual, population, community, and ecosystem. Also required is an understanding of how effects at one level of organisation create knock-on effects at other levels of organisation. For example, how changes in individuals, from genes to behaviour, translate into a change in population dynamics.The aims of the PhD will be to investigate the genetic and epigenetic mechanisms by which the model organism, Tetrahymena thermophila, responds to environmental changes. Environmental conditions such as resource availability, interspecific interactions, and temperature can be manipulated, and the effects on genes, phenotype, and population can be observed. Next generation sequencing (NGS) will provide information about genetic and epigenetic changes associated with environmental changes. Automated analysis of videos will provide phenotypes of individuals, and population dynamics.Relevant experience / expertise includes (but is not limited to):
– genetic and epigenetic analyses of microbes using NGS, including statistics and bioinformatics
– ecological consequences of epigenetic modifications
– interspecific interactions and temperature effects on microbesApplicants must have a Masters Level Degree and relevant expertise and experience. All nationalities are eligible to apply. Language of studies: English.The position will be in the group of Owen Petchey at the Institute of Evolutionary Biology and Environmental Studies at the University of Zürich, Switzerland, Petchey Group. The PhD committee will include Prof Ueli Grossniklaus & Dr Paul Hurd. Generous salary and research funding is available for at least three years. The position will be associated within the Evolution in Action research priority program at the University of Zurich. Applications should be made through the Life Sciences Zurich Graduate School , the next deadline for which is July 1st 2014 (interviews in early September). Please notify Owen when you make an application. Informal enquiries should be made to

Closing date for applications 01.07.2014

The official ad.



Interesting papers 9

Including environmental feedbacks in food web models (Sanders et al 2014).

Induced offences creates intra-guild predation (Banerji & Morin 2014).

Don’t forget ecosystem process effects on biodiversity (Lanari & Coutinho 2014).

Mice alter their foraging behaviour dependent on risk of predation (Orrock & Fletcher 2014). (Careful of why though.)

Being large and or flying leads to a long life (Healey et al 2014).

Herbivory removes negative effects of nutrients on plant diversity (Borer et al 2014).

Species loss not responsible for weak diversity-stability relationship (Hautier et al, 2014).

Browner Congo (Zhou et al 2014).

A general ecosystem model, including (nearly) every individual organism on the Earth (Harfoot et al 2014).

Higher CO2 speeds up decomposition (van Groenigen et al 2014).

Simple rules for food web stability (Neutel & Thorne 2014).


Interesting papers 8

Everything you need to know about IPMs (Merow et al 2014). Oh, well nearly everything… (Rees et al 2014).

More intuitive comparison of community composition (Baaten et al 2014).

Deriving guidelines for successful team science (Stokols et al 2008).

Algal diversity – biomass relationship consistent between experiments and lake observations (Zimmerman & Cardinale 2013).

Detecting approaching regime shifts is like detecting an approaching submarine (Carpenter et al 2014).

Moving beyond food webs, or mutualistic networks, to “inclusive networks” (Sauve et al 2014).

Connecting species- and size-spectrum food web modelling (Zhang et al 2014). But 10 fold variation in predator size unimportant for dynamics (DeLong et al 2014).

Warmer seas = smaller fish (Baudron et al 2014).

Bigger trees grow faster (Stephenson et al 2014).

Parasites promote fly diversity (Condon et al 2014).

P-values & simulation studies… NO! (White et al 2014).

Biodiversity likely has no effect on or increases infectious disease risk (Wood et al 2014).

Sheep and grasshoppers benefit each other (Zhong et al 2014).

About p-values

I stepped into a blog post by Pia Parolin titled “Do all biological processes need to be statistically significant?” ( It sounds, at moments, the frustrated cry of the field biologist observing cool patterns and building cool theories on it until he/she faces that bloody p-value=0.051. Who hasn’t been there? Yet Pia’s article contains more than that, and it raises interesting issues (also see the article’s comments). Here are some notes of mine. Continue reading ›

Some books…

Some ecology focused books you might like to read. There’s lots of variation among these, so look at the contents and aims before committing.

  • Ecology (Begon, Harper, & Townsend)
  • Community Ecology (Morin)
  • Ecological Experiments (Hairston)
  • Balance of Nature (Pimm)
  • Ecological Niches (Chase & Liebold)
  • G. Evelyn Hutchinson and the Invention of Modern Ecology (Slack)
  • Biodiversity, an Introduction (Gaston & Spicer)
  • Species diversity in space and time (Rosenzweig)
  • Nature’s services (Daily)
  • Unruly complexity (Taylor)
  • Diversity and complexity (Page)


Some statistic /methods books you might like to read. (Same advice as above.)

  • Introductory Statistics with R (Dalgaard)
  • Statistics, An introduction using R (Crawley)
  • Getting Started with R (Beckerman & Petchey)!
  • Experiments in Ecology (Underwood)
  • Observation and Ecology (Sagarin & Pauchard)
  • The ecological detective (Hilborn & Mangel)
  • Cause and correlation in biology (Shipley)
  • An R companion to applied regression (Fox & Weisberg)
  • The lady tasting tea (Salsburg)


Some notes on Principal Component Analysis

Principal Component Analysis (PCA) is an ordination method that reduces the dimensionality of multivariate data by creating few new key explanatory variables called principal components (PCs).
Each PC accounts for as much variance in the data as possible, provided that all the PAs are uncorrelated: therefore all PCs are independent and orthogonal.
It is possible to order the PCs according to the amount of total variation they explain, as well as to determine the relative contribution of each of the original variables to each PA.
A practical example follows using the software R on the “iris” dataset: Continue reading ›