There are two possible reasons why two individuals might be similar in some characteristic (e.g. two children with brown hair): (1) they are descended from some common ancestor from whom that characteristic was inherited intact, or (2) the characteristic was independently evolved in one or both ancestral lineages that resulted in the individuals. To illustrate:
(1) +———descendant 1
(2) +—X—descendant 1
bold indicates evolved character state
X indicates evolutionary change in character state
Therefore, when we want to say that two individuals are identical, it is useful to specify why: we do this by saying that they are identical by descent in the first case, and identical by state in the second. The distinction between identity by state and descent can also be couched in conventional genetic terms: two copies of an allele might share their genetic sequence because they were derived from a recent common ancestral sequence without mutation (identity by descent); or mutations have altered their sequence since divergence from a distance common ancestor, and they have recently converged on the same sequence. Put simply, identity by state and descent in genetic sequence is dependent at least on the rate of mutation and the time to common ancestry.
Identity by descent is an important attribute in evolutionary biology because it makes a statement about the ancestral history of a population, as well as its mating structure. For example, if inbreeding is historically common, then most cases of similarity will be identity by descent. This is true because in an inbreeding population, the mate of an individual will be more closely related than expected in a randomly-mating population, causing common ancestry to be more recent. In other words, two similar individuals are more likely to be cousins when inbreeding is common. The same is true when a population is divided into reproductively-insular subpopulations or demes.