We have for decision a question of whether an explosion, within the meaning of certain insurance policies, occurred in a large transformer. In answer to special interrogatories, a jury found an explosion did occur in the selector switch compartment of the transformer, causing damage, but that no such explosion had occurred in the main tank of the transformer. It is agreed that the insurance is collectible if an explosion occurred in either compartment, but the District Judge held that what occurred in the selector switch compartment, as determined by the jury’s verdict, was not such an explosion as was covered by *139the insurance in the light of the provision of the policies that “electrical arcing” is not an explosion within the meaning of the insuring clauses. Since the District Judge was also of the opinion that there was adequate evidence to support the finding of the jury that no explosion had occurred in the main tank, he entered judgment for the defendant notwithstanding the verdict.1
The large transformer, which sustained damage requiring repairs costing over $92,000, is housed in a metal casing internally subdivided into three separate compartments. The largest of these compartments, known as the main transformer compartment or the main tank, houses the cores of the three transformer phases, each with their high and low voltage windings. The selector switch compartment contains mechanical switches which control the number of turns energized in the high voltage windings. The electrical equipment in each of these compartments is immersed in insulating oil, approximately 5,000 gallons being in the main tank and a smaller quantity in the smaller selector switch compartment.
At 5:31 o’clock in the afternoon on August 1, 1956, a “power arc,” or short circuit, developed between two fixed terminals of the A-phase selector switch. Within a matter of microseconds, terrific heat developed. This heat vaporized and decomposed some of the insulating oil, converting it into gases which rapidly expanded with such pressure that the micarta diaphragm capping the opening at the top of the compartment was blown out.
The arcing across the two fixed terminals in the selector switch compartment caused similar arcing to occur across the related high voltage windings in the main tank. Substantially instantaneously, great heat was generated in the main, tank, vaporizing and decomposing insulating oil, decomposing insulation on the windings and melting some of the copper conductors. Copper particles were found imbedded in the windings of all three phases, but the jury’s finding that no explosion occurred in the main tank, for our purposes, forecloses an assumption that the distribution of the copper particles was the immediate result of an explosive force rather than the circulation of thermal currents in the insulating oil in which the copper particles had become temporarily suspended.
It was thus evident that the jury reasoned that an explosion occurred in the selector switch compartment but not in the main tank, because the micarta diaphragm on the former was ruptured while that on the latter was not. The difference seemed to the District Judge an insignificant distinction, for the diaphragm functioned as it was intended, and its loss, together with the loss of a small quantity of oil which erupted through the ruptured diaphragm, viewed as damage caused by an explosion, was trivial in comparison with the extensive damage caused directly by the electrical disturbance.
What occurred in the selector switch compartment, however, meets the usual definitions of an explosion.2 The intense heat caused a sudden conversion of liquid oil into gases which rapidly and forcefully expanded. Furthermore, the greatly increased internal pressure was suddenly released when the micarta diaphragm ruptured. Undoubtedly noise accompanied this abrupt and radical increase in pressure and its sudden release, all occurring within a fraction of a second.
*140The difficulty here arises from the fact that the policies cover damage from electrical arcing only if the arcing is followed by an explosion, or another covered peril, and the policies provide that arcing, itself, is not to be regarded as an explosion. The suggestion, therefore, that the parties must not have intended to treat as an explosion the natural and usual consequence of arcing, at first glance, has an appearance of reasonableness.
The suggestion, however, neglects the importance of the effect of the arcing upon the medium in which it occurs and upon other materials in close proximity to it. Every short circuit, perhaps, creates heat, and it may crackle noisily, but if it occurs in air, it does not usually occasion such a violent build up, or release, of pressure as to be characterized as an explosion. We deal with a very different situation, however, when a short circuit occurs in a solid or liquid medium and the generated heat is sufficient to accomplish vaporization of the medium so rapidly as to be measurable only in microseconds. If the heat of a short circuit occurring within, or near, a stick of dynamite should ignite the dynamite, it would hardly be said that an explosion had not followed the electrical disturbance within the meaning of these policies. The substantially instantaneous vaporization of a large quantity of insulating oil is not so readily achieved as that of a stick of dynamite, but, when it occurs, the violent expansion of the released gases is no less an explosion than comparable pressures of gases released by burning blasting powder.
It has been suggested that “explosion” has a more limited technical meaning, and that it should be restricted to the “rapid, exothermic, self-propagating decomposition [of explosive materials] with the formation of more stable products.” 6 Encyclopedia of Chemical Technology 1; see Hammon, An Old and a New Legal Problem: Defining “Explosion” and ‘‘Sonic Boom.” 45 A.B.A.J. 696. Only explosives can undergo such exothermic, self-propagating decomposition, and there is nothing to indicate that insulating oil is such an explosive. The definition, however, would exclude all pressure ruptures and other phenomena commonly considered to be within the protection of explosion insurance. The policies here contain no comprehensive definition of the term, but they do provide that the term shall not include concussion unless caused by explosion, electrical arcing, water hammer or the rupture or bursting of water pipes, exclusions which would have been quite unnecessary if the policies did not cover other pressure ruptures and chemical conversions of nonexplosive materials which release expanding gases with sudden and violent pressure changes. Also excepted in the policy is the rupture of specified steam and other equipment and vessels having normal internal maximum pressures exceeding designated limits relative to atmospheric pressure. Such exceptions leave little room for argument that pressure ruptures and similar phenomena were never included in the general explosion coverage.
By the terms of the policies, fire following an electrical disturbance is not within the coverage of the policies unless self-sustaining, but no words appear which limit explosions to those that are self-propagating.
The policies, as we read them, exclude coverage for damage from an electrical disturbance which does not cause, or is not followed by, an explosion or other insured peril, but do cover such damage if the electrical disturbance causes, or is followed by, a vaporization of solid or liquid materials which is sufficiently sudden and rapid, and the released pressures sufficiently violent, to meet the common understanding of an explosion.
Whatever may be said to have occurred in the main tank, the testimony warrants the jury’s finding that an explosion occurred in the selector switch compartment. The significance the jury *141attached to the rupture of the micarta diaphragm is not unjustified, for it demonstrates the presence of violent pressures released by the vaporizing oil. The diaphragm was designed to withstand pressures of 12 to 15 lbs. per square inch, so the internal pressure in the selector switch compartment, at least, approached 2 atmospheres.
The insurance companies, however, contend that the micarta diaphragm was a safety device, which functioned as it was intended, and its bursting is no more significant than the normal operation of a spring loaded valve on a domestic hot water heater.
We find no substance in the suggested analogy. Spring loaded valves and similar pressure regulators and relief devices on hot water and steam equipment are designed to release excess pressure and thus protect the equipment, but the internal pressures they release develop relatively slowly and their release of excess pressures, in normal operation, is not so abrupt as to be explosive. As safety devices, such equipment is employed to prevent bursting of the container and the explosive release of the slowly developed internal pressure. Here, in contrast, the internal pressure developed with explosive suddenness, and its release through the opening, 12-14 inches in diameter, made by the bursting diaphragm was no less explosive than its release would have been if an opening of comparable size had been blown in the side of the metal casing. Since the diaphragm was designed to withstand pressures somewhat less than those the metal casing could bear, it was a protective device intended to minimize damage from an internal explosion, but incapable of preventing such an explosion. Far from demonstrating that no explosion occurred, the fact that a diaphragm of such size and strength did break is but further demonstration of the force and violence of the suddenly developed internal pressure and the rapidity of its release.
Finally, the insurance companies, interpreting the language of the policies to cover damage from the electrical disturbance only if damage is caused by the explosion, contend that there was no such explosion damage here. The jury found there was, and we agree that the evidence warrants the finding. The damage caused by the explosion was relatively inconsequential in comparison with the cost of repairing the other damage, but nothing in the policies justifies a limitation of coverage to those losses in which explosion damage is a major or substantial part of the total. A requirement that there be some damage from explosion as a condition of coverage, may be an appropriate safeguard against claims when the occurrence is uncertain or speculative, but, concededly, there was explosion damage here which we cannot overlook merely because the other damage proved to be so extensive. Nevertheless, the insurance companies say the micarta diaphragm was intended for destruction in the event of an explosion and its loss should not be considered as explosion damage. This contention treats as negligible the loss of vaporized and erupted oil, but the contention itself is unsound. The diaphragm was intended to minimize explosion damage, but avoiding the heavier damage of loss of the metal casing at the cost of the diaphragm does not mean that destruction of the diaphragm was deliberate or was not a loss from the explosion. The explosive forces here were hostile, not friendly, and their destruction of the means by which possible greater loss was avoided, in our opinion, was damage from the insured peril.
The District Judge anticipating the possibility of our holding, has found the recoverable damage to be $92,598.00. The judgment will be reversed and the cause remanded for the entry of an appropriate judgment for the plaintiffs after such further proceedings, if any, as may appear necessary to the District Court.
Reversed and remanded.
. The relevant facts, the applicable provisions of the policies and the bases of his conclusion are carefully set forth in the opinion of the District Judge. 166 F.Supp. 703.
. Empire State Ins. Co. of Watertown, N. Y. v. Guerriero, 193 Md. 506, 69 A.2d 259; Tires, Inc. v. Travelers Fire Ins. Co., 4 Cir., 253 F.2d 411; Jersey Ins. Co. of N. Y. v. Heffron, 4 Cir., 242 F.2d 136; American Alliance Ins. Co. v. Keleket X-Ray Corp., 6 Cir., 248 F.2d 920.