3.  Non-linearity:  Linearity means that we can get a value for the whole by adding the values of its parts.   Since linearity is easier to use mathematically, considerable effort is expended to approximate linearity when describing essentially non-linear systems.  This does not work for cas.
>For example, in predator-prey modeling, we can examine complications caused by non-linearities.  Increases in either the predator or prey populations increase the likelihood of encounters between them.  The mathematics of this interaction involves a non-linearity because it entails the product of two distinct variables instead of their sum.  The overall predator-prey interaction cannot be obtained merely by adding the predator activity to the prey activity.
>These equations are a version of the Lotka-Voltera (1956) model.  Under most conditions, the predator population will go through a series of oscillations between feast and famine, as will the prey population.  
>Non-linear interactions prevent us from assigning an aggregate reaction rate to the aggregate reaction.  Non-linear interactions almost always make the behavior of the aggregate more complicated than would be predicted by summing or averaging.