Fujisawas Sake said:
Hemocyanin serves as a good transport, but has a low affinity for, oxygen. Freshwater is more dynamic in the O2 and salinity (ionic) changes than seawater. Fish, utilizing the more oxygen-efficient hemoglobin, were able to colonize freshwater first, and therefore were able to keep this niche well-stocked and defended.
Interesting thought, but it leads me to another question about evolutionary biology in general... If cephs were able to suddenly utilize a more efficient means of oxygen exchange (say, an iron-based pigment like hemoglobin, instead of the less-efficient copper-based hemocyanin), would that make a difference? If so, how many such quantum leaps in physiological changes may have led to rapid changes in evolution in the past? Could a mutation that affects something like blood-O2 affinity lead to a whole new pattern of lifestyle and/or behaviour?
When I was reading up on copper's negative effects on cephs for a thread a week or two ago, one of the articles pointed out that sometimes hemocyanin gets a bad rap for its lower O2 affinity, it has some other advantages that make it more competitive with hemoglobin than just the affinity would suggest. There are certainly plenty of freshwater and land inverts that use it for O2 transport, so it seems like in and of itself it's not a big limiter-- although most freshwater molluscs are pretty sluggish, and a lot of insects are quite fast and active... I've read conflicting things on O2 in cephs-- on the one hand, it's claimed that they're less tolerant of being temporarily deprived of O2 than hemoglobin users, but on the other some cephs (nautilus?) have some nifty mechanism to survive effectively in low-O2 conditions. Certainly, since many live in deep water with a lot less O2 saturation, there's been a lot of evolutionary pressure for those sorts of adaptations...
I've frequently been amazed at how conservative animal genomes are, once they've hit on a "successful" model... often, certain patterns that were successful in one particular niche really get locked into the genome, and a lot of diverse critters get formed by making small changes but keeping the basic patterns. That's one reason I really like thinking about cephs, because the cephalopod pattern diverged from the vertebrate one a long time ago, and at a very phenotypically simple stage, so they're really the only neurologically "advanced" animals with which we share so little history (possible exception for certain insects, but their "intelligence" seems less clear-cut to me). However, even there, the basic segmented, bilaterally symmetric body plan controlled by the homeobox genes was well-established before the divergences of insect, mollusc, and vertebrate, so we're not as different as we could be...
Anyway, I suspect that the use of hemocyanin is very well established evolutionarily, such that it's pretty unlikely that such a drastic change could occur naturally; I bet it's so entrenched in intereactions with other systems and the genes that control them that a change that major would require many other simultaneous changes to avoid being lethal just by breaking the system. I bet that there is some very strongly conserved pattern of interaction in cephs' genetic legacy that favors salt water pretty strongly, just because I can't think of any other reason why they wouldn't be around in freshwater. In fact, I know the oceans were a lot less salty when the early cephs were so dominant, so at least they could handle the freshwater-ness. I think the theory that freshwater has less of an O2 buffer is a possibility, but that's more of an issues for rivers, streams, and ponds; there are still freshwater lakes and semi-freshwater estuaries that are more ocean-like in terms of having a pretty consistent O2 level that could serve as an evolutionary stepping-stone. And heck, there are octopuses that live in tidepools, where both the salinity and O2 levels can fluctuate pretty wildly.
Since cephs were established as free swimmers so much earlier than a lot of the modern forms of life, I bet that they established a lot of genetic "weight" on the systems that were very effective for that era, and probably "locked themselves in" to some things that were selected for very strongly early on, but were sub-optimal when other critters arose. The giant axon seems like a good example of this-- when myelinated axons sped up nervous systems in fish, the squids had to improve their escape reflex, and (I'm guessing because of this genetic lock-in) rather than developing their own myelin, they just scaled up the neuron design that was successful for so long, and made it bigger for faster transport.
Hmmm.
Of course, even if there are conservative genetic reasons cephs are unlikely to have point mutations to switch to hemoglobin or freshwater or what-have-you, the "mad scientist genetic engineer"

approach could let us pick and choose things that could never occur as point mutations...