Devolution (genetically speaking) - Tree of Life loops

A complex system, while certainly effective at competing, still has the higher cost of complexity. It is not apparent to me why complexity is maintained.

It is maintained because those costs can be met.

This explains why some complexity still exists. Not why complexity does not evolve into simple forms. Why no "cycles" in the graph of life. Why don't we see any lineage that flows from low genetic complexity to higher back to lower.

Define "complexity". Many istances of character reversal and vestigiality have been noted in evolution. Is this what you are referring to?

I assume you are suggesting possible explanations for why devolution (as I described it) does not occur, but I fail to see the explanatory power in these subjects with regard to my question. The fact that simpler systems can compete is clearly evident by the success of simpler systems. A complex system, while certainly effective at competing, still has the higher cost of complexity. It is not apparent to me why complexity is maintained.

I don't know if this question fits or not in this topic, but I wonder if part of the answer lies in how much of the genome effectively impacts the organism. We have a much larger genome than bacteria, but bacteria have much less non-coding DNA. Perhaps if we looked only at the part of the mammalian genome that is expressed, it wouldn't be that much more than what you find in prokaryotes.

Dawkins suggests in The Selfish Gene that surplus DNA may be a sort of genetic parasite--"a harmless, but useless passenger, hitching a ride."

Is there a reason for it to be? Remember, there must be selective pressure on something for changes in it to be passed along.

For argument's sake, let's say 15% of our example animal is actual, honest to goodness, not coding, can be lost with no harm, truly junk DNA. The ones that do take up space. Currently noncoding is a lot more, but 15% is actually junk. What pressure would make this animal get rid of 15% of its DNA? Where would the mutation to stop coding DNA come from? How would it get only that 15% or so gone but keep the rest that is necessary? Etcetera etcetera etcetera. Remember, changes don't just randomly happen, they happen IN the DNA, as mutations or other methods. So there would have to be a change IN the DNA to make some of the DNA stop being coded, but only parts that are ACTUALLY junk, and it would have to confer enough of an advantage in the environment to be worth keeping. Since such a change hasn't arisen, we wouldn't see what you wonder about.

Good point, but you almost seem to be proposing a law that there will never be selective pressure to reduce "junk" DNA. This does not seem plausible to me. The advantage of not having junk DNA would be an energy advantage at the very least. The energy used in mitosis and transcripting useless DNA could be significant. Even if we are in an energy rich environment it still means more energy could be used in valuable biomass as opposed to useless biomass.