In general, there’s little or no correlation between the length of a eukaryotic organism’s genome and any other “obvious” characteristics, such as the creature’s overall size. Humans have a genome of about 3.4 billion base pairs; those of mice and giraffes are a bit smaller (2.7 billion base pairs), for example, and orangutans’ and guinea pigs’ are larger (about 4.0 billion base pairs) [1]. Many amphibians have gigantic genomes, 27 billion base pairs for the tiger salamander, e.g. The figure below, from the on-line version of Molecular Biology of the Cell, gives a nice glimpse of the scale of lots of organisms’ genomes. In general, genome size doesn’t mean much; it’s simply a relic of the accumulation of random bits of DNA, for example from viruses and transposons, over billions of years of evolutionary history.
It turns out, however, that there are creatures for which genome size does matter. Suppose you’re an organism with a very fast metabolism –- your heart rate is high, you consume oxygen very rapidly, etc. Your ability to function may depend on the surface area of your red blood cells, since oxygen has to pass through the red blood cell membrane to be bound by hemoglobin. You might want, therefore, lots of small red blood cells rather than fewer large ones, since ratio of surface area to volume is greater for smaller shapes. For mammals, the desire for small blood cells shouldn’t generate evolutionary preferences about genome size, since our red blood cells don’t contain any DNA. In birds, however, the red blood cells do contain DNA. Flying is a metabolically costly activity — it takes a lot of energy to fight gravity. One might expect, therefore, that birds have small genome sizes, and indeed they do: about 1.3 billion base pairs, on average, compared to the mammalian average of 3.1 billion.
Should one believe this just-so story? If true, it would predict that (i) non-flying birds have larger genomes than flying birds, and (ii) hummingbirds, with the fastest metabolism of all the birds, should have the smallest genomes of all. Both of these are, in fact, true! As discussed in a neat paper from a few years ago [2], the average hummingbird genome size is 1.0 (+/- 0.01 s.e.) billion base pairs, much less than the avian average. I’ll leave it to the reader to explore (i), either via papers on the subject, or by exploring the genome size database in [1]. (The database is fascinating.)
I stumbled on all this while thinking about shape and scaling (for the class I’m presently teaching), and while reading a fascinating draft of a book on biophysical models and wondering whether there are exceptions to the lack of sense in animal genome sizes.
Of course, if I were designing a bird, I’d remove the DNA from red blood cells entirely, as is the case in mammals. But, no one asked me, and this provides yet another good illustration that evolution selects for traits that help organisms survive (like smaller genomes for birds), but doesn’t necessarily find “optimal” configurations. All this reminds me of the excellent painting by surrealist Remedios Varo, at the top of the post.
[1] All genome sizes are from the excellent http://www.genomesize.com (and references therein). Values (“C-values”) there are picograms of total DNA; on average, 1 pg is 0.978 x 10^9 base pairs.
[2] T. R. Gregory, C. B. Andrews, J. A. McGuire, C. C. Witt, “The smallest avian genomes are found in hummingbirds,” Proc. R. Soc. B Biol. Sci. 276, 3753–3757 (2009). http://rspb.royalsocietypublishing.org/content/276/1674/3753
The Eighteenth Elephant 
Bird RBCs have an elongated shape, not sure what that means. But birds have a much more efficient air flow system so maybe there is less selective pressure to have efficient blood cells.
One interesting related story is that there is a very long gene in flies. It is so long because it has a powerful promoter that transcribes it constantly, starting from the earliest zygotic stages. But it doesn’t actually get turned into protein until later in development. The reason is that when the embryo is rapidly dividing, the gene doesn’t finish transcription because it is so long. But when the cell cycle slows, which is when it is needed, then it reaches the end.
Also, Fritillaria meleagris has a genome size of 130 Gb and can be seen in Hendricks Park, if you want to check out one of the true losers in the battle against transposable elements.
