Some thoughts on ecology, evolution and economics

Predicting cannibalism

Ecosystems are messy places.  The number of possible relationships between species grows exponentially with every species that enters the system.  A fairly typical foodweb (see image below) of 70 species has 4900 possible predator-prey interactions.  Despite that, food webs are suprisingly predictable.

food web

For one thing, most foodwebs have only 10-30% of the
possible feeding relationships.  This number is called the
connectance of the food web.  As I outlined  previously ,
if you know the number of species and the connectance
of a food web you can predict much of its structure. In
particular, you can predict:



  • the % of top species; those without predators
  • the % of intermediate species; those with both predators and prey
  • the variety in number of prey species; the connectance reflects how many species are in the average diet but this number tells us how pickiness varies across species in the ecosystem
  • the % of cannibal species; those that prey on their own species
  • the % of species in feeding loops; this number reflects species (apart form cannibals) in food chains that turn back on themselves (i.e A eats B, B eats C and C eats A). This seems odd when you think of species of very different sizes but when the species have a similar size these loops are quite common and important for maintaining ecosystem stability
  • the average # of links between two species in the food web; where one link is a pair of predator and prey species

We are not yet very good at predicting the percentage of bottom species (those without prey- usually plants or fungi), herbivores (those that eat bottom species) and omnivores (those that eat plants, herbivores and other predators in the same diet).

It is not clear why these simple models can be accurate about a range of different ecosystems.  It does suggest that ecosystems are not so messy as we believe.  Quite possibly, we have not observed enough food webs yet and we are simply fitting our models to the available data.

This work gets really exciting when it is combined with body size relationships.  Remember, the amount of food an animal eats is very closely tied to body size.  We can start to predict how ecosystems will react to the loss or introduction of  a species based on readily available information on diversity, connectance and body size.  This gives me a lot of hope about treating ecosystems as a whole rather than as a bunch of disconnected parts.



Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>