Evolution and the scabbardfish's deleting of a molecule before adding its replacement

"Imagine you're a water molecule in a glass of ice water, and you're floating right on the boundary of the ice and the water," proposes Emory University physicist Eric Weeks. "So how do you know if you're a solid or a liquid?"

 Cool.  And the article from which I copied that line is cool as a whole.




And I then referenced the linked article, and then took read a tangetal article which reads in part,

Emory University researchers have identified the first fish known to have switched from ultraviolet vision to violet vision, or the ability to see blue light. The discovery is also the first example of an animal deleting a molecule to change its visual spectrum.

 My question here is: the scabbardfish specie's precursor's having developed an ability to see ultraviolet light was certainly a fantastic development, one which only theretofore was enhanced continuously, to great benefit to that scabbardfish's ancestor.  That creature's ability to see ultravioletly was absolutely central to the specie's ability to survive, and, in all liklihood, thrive within its environment.

So, how could it be that such a valuable ability would evolve in such a way that an existent benefit would be nullified?; Of course, it happened that a gradual drift from the ultraviolet to "visible" blue light, from the creature's perspective, happened to represent, itself [the drift along the spectrum itself] an evolutionary benefit: either that creature's predators, or its prey, once seen a bit using ultraviolet light, was being rendered mre clearly visible by blue light.

But, wouldn't the existing ability to see ultraviolet light function as a pressure, acting upon natural selection, which would counter a random drift, across generations, toward a bluer sensitivity to light?

Of course.  But countering that point:

We an safely presume equally that the implied predator/prey, for the scabbardfish's ancestor, was adjusting its detectability from the ultraviolet and toward blue.


But... here, a molecule was deleted was it not?  Did the article's author truly mean that?  Was a [see ultraviolet] molecule deleted before the [see blue] molecule appear within the follil record?  Now, if that were true, well, I'm astounded.