The potential for animals to develop gas embolisms in blood and body tissues was first proposed by Robert Boyle as early as 1670. However, it wasn't until 1905 that Marsh and Gorham provided the first definitive description of bubble formation in fish and its relationship to DGS. The occurrence of GBT in fish did not receive much attention until the late 1960's when it was recognized as a serious problem in the Columbia and Snake Rivers in the United States (Coutant and Genoway 1968, Ebel 1969, 1971 and 1979, Beiningen and Ebel 1970, Bouck et al. 1970, Ebel et al. 1971, 1973, 1975, and 1979, Blahm et al. 1973 and 1975, Boyer 1974, Meekin and Turner 1974, Weitkamp 1974 and 1976, Dell et al. 1975, Dawley et al. 1976, Stroud and Nebeker 1976, Clark 1977). Large hydroelectric and impoundment dams along these rivers produced dissolved gas tensions with delta P levels approaching 400 mm Hg. At times, the losses of Pacific salmon migrating through the system of dams have been enormous. For example, Westgard (1964) reported that nearly 88% of the migrating adult chinook salmon died in the McNary spawning channel before spawning could occur. In another incidence, approximately 20,000 spring chinook salmon were lost in the vicinity of the John Day Dam in 1968 (Weitkamp and Katz 1980). The problems of DGS and its effects on fish of the Columbia and Snake Rivers resulted in a wide range of research which attempted to describe, quantify, and correct the problem. Weitkamp and Katz (1980) and Colt et al. (1986) provided extensive literature reviews on the subject.

One of the distinguishing features of earlier studies was the belief by many researchers that only dissolved nitrogen was the cause of GBT in fish. Presumably, there was the mistaken assumption that dissolved oxygen in blood could not diffuse into intra-corporeal bubbles or the swim bladder. This led to much of these data being reported in terms of nitrogen supersaturation only. As a result, these data are difficult to interpret and in many cases unusable for developing dose - response relationships.

In the late 1960's and early 1970's, DGS was also recognized as a potential problem in the Columbia River in Canada. Delta P levels approaching 350 mm Hg were measured below the Hugh Keenleyside Dam (Clark 1977). These high levels were found to persist all the way to the US border and into Lake Roosevelt. During the same period, many other rivers and lakes in British Columbia were found to have elevated dissolved gas levels (Clark 1977). Although no major DGS-related fish mortalities have been recorded for the Canadian portion of the Columbia River, significant mortalities in mountain whitefish have occurred below the Libby Dam on the Kootenay River, a tributary of the Columbia River (May 1973).

In Canada, high levels of DGS and GBT have not been restricted to British Columbia. In 1968, 1969, and 1972, GBT was identified as the cause of large mortalities in Atlantic salmon and eels below the hydroelectric generating dam on the Mactaquac River of New Brunswick (MacDonald and Hyatt 1973, Penney 1987). In Manitoba, during the stocking of frozen lakes with rainbow trout, it was found that DGS caused significant levels of mortality before the lake ice melted (Lark et al. 1979, Mathias and Barica 1985).

In the early 1980's, it was becoming evident that DGS was also a problem in the shallow water rearing environments of fish hatcheries (Jensen 1980, Wright and McLean 1985, Krise and Herman 1989). Because most of the research was devoted to Pacific salmon, the problem was first identified in salmon hatcheries of the Northwest United States and British Columbia. In most cases, the causes of DGS were of natural origin, involving solar heating of hatchery water sources such as lakes and rivers (Wright and McLean 1985) or ground water which had become supersaturated (MacKinlay 1984, Miller et al. 1987). In the hatchery environment it was noticed that the signs of GBT were appearing at dissolved gas levels well below those observed in adult fish in the Columbia and Snake Rivers. Wright and McLean (1985) reported that, over a 122 day exposure period, chinook salmon fry experienced mortalities of about 4.1% at delta P levels ranging from 0 to 46 mm Hg in a rearing environment less than one metre deep. Control animals suffered a mortality level of about 1.6%.

In the mid 1980's, Alderdice and Jensen (1985a and b), Jensen et al. (1985 and 1986), Schnute and Jensen (1986), and Jensen (1988), of the Department of Fisheries and Oceans' Pacific Biological Station, conducted a variety of experimental studies and statistical data analysis of GBT in fish. The results of their work confirmed that low levels of TGP could be harmful to fish. Alderdice and Jensen (1985b) were also able to explain the high resistance of salmon eggs to the effects of DGS. They showed that the high capsule (zona radiata) pressure prevented bubble growth at delta P levels above those which normally caused bubble growth in juvenile and adult fish.

At the University of British Columbia, Fidler (1984 and 1988) and Shrimpton et al. (1990a and b) conducted theoretical and experimental studies of the physical and physiological causes of GBT in fish. The focus of this research was the definition of dissolved gas thresholds associated with specific signs of GBT in fish and the observation of the behavioural responses to these thresholds. Their results confirmed the existence of three distinct dissolved gas thresholds for certain signs of GBT. The lowest threshold (about 25 mm Hg) was for swim bladder over-inflation and accompanying over-buoyancy in juvenile fish. A second threshold existed at a delta P of about 76 mm Hg and was associated with extra-corporeal bubble growth in gill lamella of adult fish. This was also the threshold at which sub-dermal emphysema occurred. The highest threshold, corresponding to a delta P of about 115 mm Hg, was associated with bubble growth in the cardiovascular system.

In the late 1980's, Krise and Herman (1989), Krise et al. (1990), and Krise and Smith (1991) began studying the effects of low levels of DGS on hatchery-reared lake trout and Atlantic salmon. Their results showed that, like other trout and Pacific salmon species, larval life stages of these fish were fairly resistant to the effects of DGS. However, this resistance decreased as fish grew. Their results indicated that for a given fish length, lake trout were more resistant to the effects of DGS than other trout and Pacific salmon species. They also reported that Atlantic salmon were less tolerant of DGS than lake trout.

Starting in 1985, the US Bureau of Reclamation, in conjunction with Montana State University and the Montana Department of Fish and Game began an extensive experimental program on the Bighorn River in Montana (White et al. 1991). This river, like the Columbia River, experienced very high levels of DGS. The Bighorn River was distinct in that it was very shallow and had large components of DGS caused by a dam, solar heating, and primary production. Over a four-year period, these studies, involving both field and laboratory experiments, examined virtually every aspect of DGS and GBT in the river, including most river aquatic communities (i.e., all life stages of fish and invertebrates along with predator and prey relationships).