A little documentation of the alleged behavior, if you please.
This sounds suspiciously like another one of those cases where someone gete the bit in his teeth and makes highly speculative leaps of logic, without much, if any, proof, and then draws a gross generalization from it.
A quick search of the internet seems to show
1. The worms are parasitic on any number of species (including, rarely, humans), which happen by water--and every species has to ingest water somehow, so at some point most species are near water, so are potential hosts (so arguing it is a specific grasshopper-worm pair ID is wrong right there).
2. Since the statement is made that the worm larvae that don't find water die, apparently the host doesn't always get to water before it's eaten up inside. We are given no idea of how many hosts die near water. It may be only one in a hundred, or some low number. We don't know. Since only those hosts that die near water let the worms develop to their next stage, kind of by definition those are the ones that are examined, but even by pure chance some of the host species will die near water or in water, whether infected or not.
The whole suicide protein thing seems highly speculative. Where is the research on it? The only thing I found on the web in the first thirty or so hits that MIGHT deal with that was on a blog with restricted access, so I'm just going by the google summary. And a blog is certainly not by any means necessarily a knowledgeable source.
So if you're gonna argue hair worms, give us some REPUTABLE facts.
http://www.uvm.edu/extension/publications/el/el222.htm
Life History: Depending upon the species, breeding--often in writhing masses of a dozen or more worms--takes place in spring to early summer or fall. The eggs are deposited in the water in long, gelatinous strings, which swell and sometimes exceed the size of the female, containing several million eggs. Adults die after mating and egg laying. The eggs hatch in 15 to 80 days, depending upon temperature. The larva is unlike the adult, having a protrusible proboscis armed with spines. It has a very brief free-swimming existence as the larvae usually encyst on the surface of vegetation, etc., at or near the water's edge within 24 hours of hatching. The cysts remain viable for a month or more on moist vegetation adjacent to or in the water.
When a grasshopper, cricket, beetle, or other plant-eating, chewing insect ingests the cysts with the vegetation, the cyst walls dissolve and the larva bores into the host (hemocoel). Within the host, the larva digests and absorbs surrounding tissues as nourishment. The larva grows and develops into the adult hairworm in several weeks to months. At that time, it is a tightly curled mass within the insect. Sometimes there is more than 1 worm per host, and some overwinter in the host body. If the insect becomes wet or falls in the water at this time, the mature worm bores out through the body wall and becomes free living.
Hosts: Crickets, grasshoppers, and terrestrial and aquatic beetles are the normal hosts. However, the hairworms appear to be non-specific parasites and a few have been reported emerging from caddisflies and dragonflies. It is possible that predacious aquatic insects ingest the larva with their prey. The marine hairworms parasitize hermit crabs and true crabs. Recently, leeches have also been reported as hosts.
There are some records of human parasitism. It is believed that this has been the result of accidental ingestion of the worms in food or water. Hairworms are not considered to be human parasites.
Control: Hairworms are often reported on turf areas around the edge of swimming pools, etc. Reducing the number of insects attracted to and falling into the pool should reduce the hairworm problem.
In any event, the period of annoyance should be of short duration. There are no known treatments for infested pet crickets, diving beetles, etc.
--------------------------------------------------------------------------------
