Genome size varies not by a few kb worth of genes but by orders of magnitude
between organisms which can otherwise appear very similar. Whole books have
been written on this topic (see for example, Cavalier-Smith, The Evolution of
Genome Size, John Wiley & Sons ltd, 1985). The idea of whether this extra DNA
has a function is still debated. But we do know that most of this extra DNA
is not genes (we know this from old hybridization experiments and from the
genome sequencing projects, which are now giving us a glimpse at the gene
density & organization in widely varying organisms such as yeast and human).
As for a function of the extra DNA, there are basically two schools of thought: first, that the genomes of some organisms have collected vast amounts
of excess DNA (mostly nongenic) and that this DNA is completely functionless
and not tremendously deleterious. Second, that certain organisms allow their
genomes to amass excess DNA for reasons of cell cycle control (more DNA slows
S-phase and thus cell cycles). One could imagine that there are many forces
which allow genomes to expand: polyploidy, uneven crossover, transposition, etc.
Given that, if DNA replication is a limiting factor (for reasons of resources
or generation time or whatever), then these sequences are effectively eliminated
from the genome. If this is not such a priority for the cell, then these
new bits of DNA are allowed to flourish. Certain types of repeated sequences
seem to have invaded and expanded in certain genomes very quickly, like the
Alu elements in human DNA. There are also strange correlations between
genome size and things like latitude (for plant species).
