Bored researchers can think of anything.
Harvard University has a laboratory where researchers breed and study mutant zebrafish with the intent of finding the genes involved in the development of the retina. Innocuous enough research, presumably, and it receives ample funding from the appropriate federal agencies. But any researcher is bound to get bored after years of studying the same thing, and somebody a few years ago came up with the idea of using the lab's readily available livestock and equipment to study the hereditary basis of addiction. So they procured some high-quality cocaine, presumably through legitimate channels, and proceeded to give a few lucky fish the time of their short, constricted lives.
What they used:
Second-generation descendants of zebrafish that had been exposed to a chemical mutagen, along with a bunch of "normal" fish, were tested on three pieces of apparatus:
The divided tank
Apparently the best way to deduce the effects of a habit-forming substance in lower vertebrates is to use a technique called "conditioned place preference." The fish enter a rectangular tank, split into equal volumes by a partition with holes big enough for the fish to swim through. A fish's natural "preference" is then determined by the percentage of time it spends on each side of the divider. Each fish is then blocked off in its less preferred side and its aqueous environment is dosed with a generous infusion of cocaine. Most get a real kick out of this. (But those that don't are the interesting ones...)
The following day, after the fish have had ample time to come down off their high, they're readmitted to the now-denarcotized tank, its central partition once again fish-permeable. There's no dope to be found on either side, but the great majority of the fish in both mutant and control populations have developed a newfound preference for the formerly drug-filled territory. No noticeable difference, however, is seen when the zebrafish have been exposed to the non-"rewarding" anesthetic lidocaine rather than the pleasure-inducing cocaine.
The vision test
The fish then go in a bottle surrounded by a rotating cylinder, painted white on the inside with a black spot. When a normal fish sees the spot rotate into view, it hides behind a post in the bottle. But, of course, it has to see the spot first. The fish's "visual threshold" is determined by the minimum amount of light needed for it to see the spot and get out of the way. Cocaine tends to make the fish less sensitive to light; lidocaine doesn't.
The maze
Maybe the fish that didn't notice the cocaine have some sort of mental malfunction. To test for posssible stupidity, the experimental subjects go into a maze, and the time it takes for them to learn to swim to a "favorable habitat" is measured. No, they don't get any drugs this time around.
What's wrong with these fish?
However amusing it might be to watch a doped-up zebrafish swim around, the significant ones from a genetic standpoint are in the minority that doesn't seem to be influenced by the drug. And particularly when your funding comes from a government that considers unprescribed psychoactive chemicals a priori bad, you really should be studying those fish that might make good role models for impressionable D.A.R.E. youngsters.
Three out of eighteen of the second-generation mutagen-exposed zebrafish families had a large percent of members that just didn't seem to have any interest in cocaine. None of the unmutagenized groups did. Drug-indifferent fish from the three families were inbred, and their offspring were tested once again, along with the offspring of their normal siblings, which were used as a second control. The three groups were:
Dumbfish
(dum)
So named because they were, well, dumb. Cognitively impaired, you might say. While unresponsive to cocaine, the dumbfish had extreme difficulty navigating the maze that their cocaine-susceptible cousins could learn relatively quickly. Many of the dum group never learned the way, even after having been chased through it repeatedly by the experimenters. And the drug didn't significantly alter the fishes' visual threshold, either. No mention of whether it was below average to begin with, though.
Jumpy
(jpy)
These fish, for the most part, didn't have any great difficulty learning the maze per se, but they had plenty of other troubles. They were skittish and easily stressed, flipping and flopping and rolling over, not swimming straight, and generally overreacting to unfamiliar situations. Many behaved so weirdly that they simply couldn't be tested. Their lack of preference for cocaine was apparent, though, as was that drug's negligible effect on their visual acuity. Maybe a stimulant just cant do very much for someone/thing that's already perpetually on edge.
Goody-two-shoes
(gts) (I am not making these names up!)
Unlike their rather distant cousins in the dum and jpy families, the fish of gts type were clearly physiologically affected by the narcotic. It impaired their eyesight just as much as it did for the "normal" fish. But for some reason, the majority of the goody-two-shoes zebrafish simply showed no interest in cocaine. In fact, they showed the lowest average percent increase in preference of any group tested. And they weren't stupid, or neurotic, or psychologically screwed up in any other perceptible way. Apparently, their brains aren't wired to obtain any sort of stimulation from the drug in question.
For further reading, see the original article by Darland and Dowling in Proceedings of the National Academy of Sciences of the United States of America, vol. 98, no. 20, pp.11691-96 (September 2001). It's apparently also available on the Web at www.pnas.org/cgi/doi/10.1073/pnas.191380698
Harvard's cocaine-proof zebrafish continue to be bred and studied. There's no mention of what happened to the leftover cocaine. | 
