1. Conclusion in brief

The documentary record supports four propositions, but not a fifth:

• • an assorted stockpile of naval ordnance still containing explosives lay on the beach;
  • the devices had been inspected and stripped of their normal initiating mechanisms, but had not been emptied;
  • the Allied military inquiry concluded that the ammunition had been deliberately detonated by a person or persons unknown;
  • the material remained accessible, unguarded and inadequately isolated in a place frequented by the public;
  • what has not been established is who physically prepared the initiation system, or who may have issued the order.

The OZNA lead is historically plausible and is supported by a British intelligence report dated 19 December 1946. That paper, however, records information originating with Italian counter-intelligence and is neither the conclusion of the technical inquiry nor judicial proof.¹

2. The essential context

In August 1946, Pola belonged to Zone A of Venezia Giulia, administered by Anglo-American forces under the Belgrade Agreement of 9 June 1945. Most of the remainder of Istria lay in Zone B, under Yugoslav military administration. The city’s territorial future was being discussed at the Peace Conference, while organisations supporting Italy and organisations supporting Yugoslavia confronted one another within Pola’s society.²

The Paris Peace Conference met from 29 July to 15 October 1946 to examine the treaties with Italy and Germany’s other former European allies. For Italy, one of the principal issues was the delimitation of the border with Yugoslavia in Venezia Giulia.

3. A minimum chronology of 18 August 1946

Contemporary sources differ over the visible details of the blast, but agree on its suddenness, the concentration of victims among families close to the stockpile and the initial difficulty of identifying many remains.⁵

4. How many victims were there?

The number of victims has itself become a subject of dispute. Figures of 100, 110 or 116 deaths are repeated in memoirs and public discourse, but have no demonstrated correspondence in nominal lists, burial registers or lists of missing persons.

Čonč concludes that the overall number should not significantly exceed seventy: the number of remains buried substantially corresponds to the known names or reported missing persons. A few unregistered victims cannot be excluded, but this does not justify turning 63–65 documented deaths into 110–116.⁶

Cautious formulation: at least 63–65 documented deaths; a small additional number is possible; a figure exceeding one hundred has not been demonstrated. The new memorial in Pola records 64 identified victims.⁷

5. What exploded?

The formula “twenty-eight Italian anti-landing mines containing nine tonnes of TNT” does not match the inventory emerging from the British reports. The sources describe a heterogeneous stockpile; moreover, the number of objects present and the number that actually detonated need not be the same.

Dukovski expressly notes the divergence among the documents: two refer to 15–20 depth charges, while a third, produced between September 1946 and February 1947, lists three torpedo warheads, four TNT charges and five smoke generators.⁸

Mines, depth charges and torpedo warheads had different purposes and initiating systems, but could contain substantial masses of high explosive. It is inaccurate to describe their contents generically as “highly compressed TNT”: depending on model, origin and period, cast TNT or different explosive mixtures could be used. The stockpile’s heterogeneous composition also requires a distinction between the total quantity present, the quantity actually detonated and possible sympathetic propagation from one charge to another.

The ordnance had been recovered from the waters of Pola during clearance operations. The sources do not permit its manufacture to be attributed uniformly and exclusively to Italy: both German and Italian material is mentioned. In legal-administrative terms, Croatian sources maintain that Yugoslavia claimed the material as war booty, although it remained under Allied control pending the competent decisions.⁹

6. “Disarmed” does not mean empty of explosives

The inspections were intended to render the material incapable of exploding under ordinary conditions. There is no indication that the explosive charges had been removed. A device without its normal fuze may still detonate if fitted with a new initiator or struck by an adequate detonating charge.

A summary of the Venezia Giulia Police Force investigation, updated at noon on 24 August 1946, reports that twenty-nine statements were taken and three experts consulted: a Royal Engineers officer, the head of the bomb-disposal squad and an independent engineer. According to the report, the devices could not have detonated without the insertion of a primary charge and the use of a fuse or equivalent device. The police conclusion was explicit: “These mines did not go off by accident. They were wilfully set off by the insertion of a primary charge and the using of a fuse.” This is a technical and investigative finding concerning the deliberate origin of the explosion, not an identification of the perpetrator.¹⁹

Trinitrotoluene, commonly known as TNT, is a secondary explosive relatively insensitive to ordinary impact, friction and flame when compared with the primary explosives used in detonators.

• • TNT melts at about 80.4 °C and, if ignited in limited, unconfined quantities, will normally tend to burn rather than pass immediately to detonation.
  • Detonation normally requires a sufficiently intense shock impulse supplied by a detonator and, depending on the charge configuration, a booster. The detonation velocity of high-density TNT is approximately 6.9 km/s, although it varies with density and charge conditions.
  • The absence of ordinary fuzes greatly reduced the possibility of accidental explosion, but did not make the devices harmless or absolutely inert. Intense heating, confinement, contamination, initiating components that had not been removed, or the detonation of another charge could produce more complex effects, including sympathetic detonation.
  • It would not have been necessary to reinitiate every device separately: a sufficiently energetic initial detonation could have propagated to adjacent charges.

A mere reference to a match, cooking brazier or small surface fire is therefore not, by itself, a technically adequate explanation for the almost simultaneous detonation of a large mass of ordnance. Deliberate initiation is compatible with the event’s characteristics and with the Allied inquiry’s conclusions; the device actually used, however, was neither recovered nor identified.¹⁰

Allied documentation records an inspection carried out on 27 July 1946, approximately three weeks before the massacre, and the certification of the material as disarmed and safe. This circumstance supports the sabotage hypothesis, but makes the decision to leave a large quantity of usable explosive on the site still more questionable.¹¹

7. Custody and responsibility of the Allied Military Government

The stockpile lay within the area administered by the AMG. The Allied authorities knew of its existence and had ordered or accepted the inspections. The material was not permanently guarded, was not effectively isolated from the public, and bathers could approach the devices.

The ordnance remained on the shoreline for a prolonged period, awaiting removal, allocation or final destruction. No previously ordered plan for “disposal at sea” is documented with sufficient precision. Whatever the origin of the explosion, storage in an accessible area demonstrates inadequate segregation, surveillance and custody. To classify this as a breach of a specific operational rule, however, the order or regulation then in force would have to be identified.

The administrative file also preserves a resolution adopted by the Municipality of Pola on 19 August 1946 and forwarded to Colonel Bowman on 22 August. The Municipality denounced the “absolute lack of vigilance” and stated that, since October 1945 and on several occasions, it had alerted the Allied authorities to large quantities of explosive material left unguarded, also receiving written assurances that measures would be taken. The document proves that this formal complaint was made on the day after the massacre. The file examined here, however, does not contain the complete series of earlier notifications cited by the Municipality and therefore does not allow their individual dates, recipients and replies to be verified.²⁰

Erskine acknowledged that the material was neither fenced nor guarded because it was considered harmless. In his closing report of 14 February 1947, Brigadier Bastin argued that posting guards was impracticable and that no blame attached to the Allied authorities.¹²

That was an internal assessment by the administration concerned, not a decision by an independent body. Even if the detonation was the work of saboteurs, an administrative responsibility for prevention remains: criminal interference with accessible explosives was precisely one of the risks that custody was meant to prevent.

8. What the Allied inquiry concluded

Erskine’s decisive sentence was: “The ammunition was deliberately exploded by person or persons unknown.” His reasoning referred to the earlier inspection, the impossibility of normal self-detonation and testimony concerning a small preliminary report and the beginning of a fire immediately before the main blast.¹³

The wording establishes, in the commission’s view, deliberate causation. It does not establish the perpetrators’ identity, organisation, nationality or motive. The later information concerning Kovacich was not incorporated into a judicial conclusion and did not formally alter the finding of “persons unknown”.

Police documentation shows that at least three individual leads were investigated. One suspect was excluded because he had been absent from Zone A for about two months; Antonio Zannocchi was arrested, questioned and subsequently released after his alibi was checked; a third person, described by witnesses and by then having left Pola, was sought without evidence emerging that could identify him as a perpetrator. The investigations, therefore, were not non-existent, but they produced no proven individual attribution.²¹

9. The “Sabotage in Pola” document

The report of 19 December 1946 states that a counter-intelligence source considered reliable named Giuseppe Kovacich as “one of the saboteurs”. It described him as a specialist in terrorist operations, an OZNA courier between Fiume and Trieste, and as not having been seen in Venezia Giulia since the explosion.¹⁴

The expression “reliable source” concerns an intelligence assessment, not judicial proof of the report’s content. The British report attributes the information to a counter-intelligence source; later literature has connected its provenance to Italian military intelligence circles and, in particular, the 808th Counter-Intelligence Battalion. This connection must be presented as a reconstruction of provenance, not as the complete and explicit identification of the original source within the document. Identity, means of acquisition and independent corroboration remain incomplete. The report must also be read within the political and intelligence conflict between Italian and Yugoslav agencies in the early post-war period.¹⁵

10. The OZNA hypothesis: supporting elements and limitations

OZNA, an acronym for Odjeljenje za Zaštitu Naroda — “Department for the Protection of the People” — was communist Yugoslavia’s political and security service. The Latinised Serbian spelling is also Odeljenje; Odjeljenje is the form normally reflected in the historical acronym.

A further negative but documentarily significant fact emerges from the now fully accessible Allied Control Commission file: its seventy-eight frames — comprising police reports, interrogations, administrative correspondence, municipal resolutions and relief records — contain no attribution of the massacre to OZNA. This absence does not demonstrate that the Yugoslav organisation was uninvolved; it does demonstrate that this Allied file cannot be used as documentary proof of its responsibility.²²

The correct conclusion is not “OZNA was innocent”, but “OZNA’s responsibility has not been demonstrated”. Capability, opportunity and possible motive make it a serious lead; they do not replace proof of the concrete act.

11. Alternative hypotheses

Buich’s testimony illustrates the methodological problem: he claimed to have seen Riccardo Saccon throw a lit brazier into an open mine, but the account appeared publicly many decades later and lacks confirmation. Čonč herself points to its inconsistencies and the narrator’s political position.¹⁶

Gaetano Dato keeps three leads open — Yugoslav, Italian monarchist-fascist and Yugoslav anti-communist — stating that he personally regards the first as more probable, while acknowledging that the other two are unproven and that even the Yugoslav lead falls short of definitive demonstration.¹⁷

12. Four linguistic areas, four memories

The available Slovenian sources primarily treat Vergarolla as an element in the memory of the exodus. Miha Kosmač observes that the explosion is frequently invoked as proof of Yugoslav violence and as a cause of Pola’s abandonment, but offers no independent demonstration of OZNA responsibility.¹⁸

13. Vergarolla and the exodus from Pola

The historical connection must be formulated without confusing chronology and causation. Before the massacre, declarations of intent to leave were already being collected and political mobilisation against transfer to Yugoslavia was already under way. After 18 August, Vergarolla became for many inhabitants of Pola a concrete representation of future insecurity.

To say that the massacre “caused the exodus” erases international negotiations, the change of sovereignty, fear of the communist regime, citizenship options, Italian propaganda and family dynamics. To say it had no effect is to ignore testimony and the sequence of events.

14. Final evidence matrix

15. Conclusion

Vergarolla is not a solved crime. British documentation makes the hypothesis of deliberate action stronger, but does not identify the perpetrators. The report concerning Giuseppe Kovacich makes the OZNA lead worthy of investigation; it does not turn it into a definitive judicial or historiographical truth.

Two forms of responsibility must therefore remain distinct. The first, criminal responsibility, belongs to persons still unknown. The second, administrative responsibility, concerns a military government that allowed a large quantity of explosive material to remain accessible in a city marked by political confrontation, clandestine agencies and international tension.

The victims’ memory does not require a manufactured verdict. It requires precisely the opposite: that an intelligence source not be called proof, that an estimate not be called a census, and that the absence of an identified culprit not become a pretext for erasing what the documents do allow us to state.

Alessandro Gombač

Principal sources and further reading

• Allied Control Commission, Region No. 13 – Venezia Giulia and Trieste Province, Administration, file 216, “Vergarola Explosion”, August 1946–December 1947, 78 frames, ACS reproduction from NARA.
• U.S. Department of State, FRUS 1946, report of the Commission of Experts on the Italo-Yugoslav boundary, 27 April 1946.
• Tea Čonč, “Eksplozija na Vergaroli u Puli 18. kolovoza 1946. godine”, Povijest u nastavi, 2009, pp. 133–176.
• Darko Dukovski, “Povijesna ekspertiza tragedije na pulskoj Vargaroli”, Histria, 2011, pp. 79–112.
• Istrapedia, “Vargarola (Vergarola)”, updated summary and bibliography.
• Miha Kosmač, “Organizirana izselitev prebivalstva iz Pulja”, Acta Histriae, 2015, pp. 511–530.
• Gaetano Dato, Vergarolla, 18 agosto 1946, LEG, Gorizia, 2014 — record/review.
• Glas Istre, “Vergarola – žrtve između dva žrvnja”, 18 August 2019.
• NIOSH Pocket Guide to Chemical Hazards, “2,4,6-Trinitrotoluene”, physical, chemical and safety data.