The world's scientific and social network for malaria professionals
Subscribe to free Newsletter | 11113 malaria professionals are enjoying the free benefits of MalariaWorld today

Learning about density dependence the hard way

May 23, 2010 - 21:06 -- Mark Benedict

What do data analysis and hitting mules in the head with a board have in common? Too much. Ouch!

Many MalariaWorld readers will not know this old joke, so I’ll set up the rest of the article with it.


A priest was walking down a dirt road one day and came upon a man with a mule pulling a wagon load of wooden boards. The mule had stopped and refused to move. The man was yelling, cussing and pulling on the mules' reigns trying to get him budge. The priest came up to the man and said, "If you want one of God's creatures to do what you want him to do, you must treat him with love and kindness."

Disgusted, the man told the priest, "If you can do any better, then you are welcome to try."

So the priest went to the rear of the wagon, pulled off a board and walked to the front of the mule and looked the mule sternly in the eye and then took a mighty swing with the board, hitting the mule across the head.

The man was shocked and said, "I thought you must treat him with love and kindness?"

The priest said, "You are, but first you must get his attention."


Sometimes, the realities of data that we analyze ourselves are the boards that get our attention and change our perspective on work that we are doing. This came to mind recently in regard to the effects of larval control on adult population sizes. Wow! Jumping from hitting mules with boards to mosquito control is a big leap. Indeed. Stay with me.

Concern has been expressed in opinion and scientific articles about the likelihood that negative density dependent survival of mosquito larvae could mitigate the effects of incomplete control of larvae in breeding sites. This can result from any control method including genetic, predators, biocontrol etc. In fact the ONLY larval controls excluded from this concern would be those that reliably and completely kill all larvae in each habitat.

To illustrate the problem using an example of classical sterile insect technique, imagine a habitat in which 100 eggs hatch. The carrying capacity of the habitat may be only 10 larvae, so roughly 10 larvae will survive to adulthood. If we conduct a genetic “control program” that sterilizes 90% of the females and 100 eggs are laid in the same habitat, 10 larvae will survive. (Even worse is the prospect of overcompensation in which reducing the number of larvae results in a higher total number of survivors! Worse than no effect described in the first example, “control” in this case actually increases the population. Perhaps I’ll consider this possibility in a future blog.)

The scenario I have described might only rarely exist: perhaps the number of hatching eggs approximates the carrying capacity of the habitats. But maybe not.

My head was hit with the figurative board when looking at (submitted) data that Rosemary Lees and Jeremie Gilles generated at the Int. Atomic Energy Agency. Not to give much away, they demonstrated very clear examples of negative density dependence in lab studies of Anopheles arabiensis. This is by no means the first time it has been observed, but poring over it, graphing it, thinking about it, hit my head like a board. It also caused me to think more carefully about demonstrations of effects of genetic control experiments and field projects. Providing larval competition by any means in a genetic control program (e.g. RIDL technology) could be key to success. Any system that promotes competition in larval habitats or combines genetic control with another population suppression method will clearly have a better chance of success.

We all learn in different ways. A presentation that impresses one can seem bland and academic to another. For me, getting my hands and head wrapped around a good data set drives knowledge deep into my mind like no other activity or presentation can. Sometimes, we scientists need to be hit in the head with a board before we can hear the message.



William Jobin's picture
Submitted by William Jobin on

Good message Mark,

And the joke wasn't bad either. But in Illinois, the joke is about a bride who rides off with her new husband, and has trouble with the mule. I digress, unless you want to hear my Illinois mule joke.

You have identified one of the two major difficulties in larval control, and one of the reasons why adult mosquito control by indoor spraying has been more successful against anophelines. The other reason is the difficulty in managing the people who do the spraying or other treatment of the larval habitats, in the scattered pools where anophelines breed. You need supervisors upon supervisors to make sure that people don't just dump their insecticide and go for a quiet beer somewhere.

But another ray of hope is that you can overcome this density dependence problem with a shovel, or a ditching spade, instead of a board! When you reduce the habitat extent, you cause a proportional reduction in larvae numbers.

You cited something like 10 larvae per habitat. For discussion, let us say that is 10 larvae per square meter of habitat. So reducing the total habitat from 1 million square meters to 1 thousand square meters will give you a 99.9% reduction in habitat area, and also a 99.9% reduction in larvae.

So sharpen your spades.


William Jobin Director of Blue Nile Associates