Early vision was colourful

UPDATED:

To give some of my colleagues at the University of Queensland some link love, it is being reported that they have sequenced the Queensland lungfish (currently under threat by a proposed dam) opsin genes, showing that they see in ultraviolet and visible light, as well as having the ability to see in dim and bright light. The paper is now accessible at BMC Evolutionary Biology.

The conclusion drawn from this is that early land dwelling vertebrates saw in colour, which is probably true, but not, I think, because the lungfish is a "living fossil that dates back 400 million years". It is a modern species that is the last example of a group that dates back that long. So it may be that it has derived colour vision. However, if the sequences are similar (homologous) to those in other vertebrates, the most parsimonious explanation is they both get these genes from a common ancestor.

It's also worth noting that our ancestors probably had only two receptors, one of which duplicated to give us green light reception. Humans have a very poor colour spectrum compared to some others - say the mantis shrimp, which has 7 receptors that evenly cover the spectrum from visible to ultraviolet. We, on the other hand have two receptors that haven't differentiated much from each other:

Human vision

Mantis shrimp (stomatapod) vision