(dissenting).
I am unable to agree with division II and the result.
No dispute exists that the “hazard” which Midwest was penalized for failing to guard against was the explosion of October 4, 1979. This explosion was not a “recognized hazard” within the meaning of Iowa *406Code section 88.4 (1979) unless an explosion of that nature was foreseeable. See Bethlehem Steel Corp. v. Occupational Safety and Health Review Commission, 607 F.2d 871, 874 (3rd Cir.1979). At most the record in this case contains substantial evidence that a violent phenomenon of a different nature, called an eruption, might occur. The hearing officer recognized this distinction and accordingly found that no violation had been shown. In disagreeing with the hearing officer, the review commission held it was sufficient for the bureau of labor to show that Midwest should have known of the hazard of an explosion of “any nature.” The effect of the commission’s decision was to make Midwest responsible for recognizing the hazard of the October 4, 1979, explosion despite the absence of any evidence of foreseeability of an explosion of that nature. I do not believe a violation of section 88.4 is established in these circumstances.
As used in this record, the term “eruption” refers to the sudden release of trapped gas through the furnace burden or crust. The record shows eruptions routinely occur in submerged electric arc carbide furnaces of the kind involved here. Water leaks in the furnace cooling system increase the chances of eruptions because water contributes to furnace gases and reduces the porosity of the charge bed. One expert explained the reason for eruptions as follows:
Submerged arc furnace, the attempt is to operate it in such a manner that the high temperature gases which are generated by the actual chemical reactions which [are] occurring essentially at the tips of the electrodes will filter up through the incoming charge and preheat the charge before it actually gets into the reaction zone underneath the electrode.
You’re only partially successful at doing this because the charge does not flow freely. It tends to pack from time to time and when this occurs the gases have to find their way out somewhere; so they form a channel, directly blow out and [very] hot, high temperature gases come directly out of the furnace, from the reaction part of the furnace.
Eruptions are recognized hazards in submerged arc furnaces. Precautions exist to minimize them, including shutting down the furnace when they become severe. Eruptions were occurring in the furnace involved here within the twenty-four hours prior to the explosion. Routine safety measures were employed, but the furnace had not been shut-down at the time of the explosion.
Even if eruptions are characterized as a form of explosion, however, it is uncontro-verted that the event of October 4, 1979, was not an eruption. Eruptions occur at the surface of the furnace charge and by their nature cannot occur elsewhere. The event here was initiated deep in the furnace crucible beneath the electrodes. Although various experts speculated about possible causes, no one tied the explosion to the eruptions or offered any testimony from which Midwest or anyone in the industry could be charged with knowing such an explosion might occur.
This is not a case where small explosions of the same nature had previously occurred and the accident simply resulted from a larger explosion. The ease is therefore distinguishable from Titanium Metals Corp. of America v. Usery, 579 F.2d 536 (9th Cir.1978), in which smaller explosions of the same nature had previously occurred. The case is also distinguishable from Kelly Springfield Tire Co., Inc. v. Donovan, 729 F.2d 317 (5th Cir.1984), where expert testimony was adduced to show that an explosion was a foreseeable consequence of the company’s operation of its dust collection system. No such testimony was received in the present case. Moreover, the court’s reliance on Bethlehem Steel Corp. is misplaced. It is true a recognized hazard may exist that is unrelated to the accident which preceded the inspection. See 607 F.2d at 874. In this case, however, the bureau of labor asserted that the explosion which caused the accident constituted the recognized hazard. In these circumstances the possible existence *407of an independent hazard, neither alleged nor proved, is not relevant.
I do not believe Midwest should be penalized for maintaining a recognized hazard in the absence of evidence that an explosion of the nature that occurred here was foreseeable. Evidence that explosions or phenomena of a different nature were foreseeable does not suffice.