The first public sitting · Evidence, accountability & reform
What the Anse Cochon hearing tells us.
And what it does not.
Six witnesses described omitted precautions, competing accounts of the collision, and a gap between the captain named on paperwork and the people operating a vessel. The testimony is substantial. Responsibility is not yet settled.
On this pageStart here
01 / Three-minute orientation
The hearing put serious questions on the public record.
The strongest reporting starts with what was actually acknowledged, what witnesses describe, and where the evidence still stops.
On the night of 18 July 2026, two vessels collided near Anse Cochon on Saint Lucia’s west coast. Chair Justice Mario Michel opened the public sitting by describing four deaths, injuries, the destruction of one boat and the sinking of the other. The inquiry’s remit extends beyond the collision to maritime law, operating procedures and safety recommendations.
The most revealing part of the first day was not a completed account of the impact. It was the accumulation of specific statements about how the voyages were prepared and operated: lights off during Ocean Angel’s departure, no safety briefings, jackets not routinely worn or distributed, an accepted operator-licence expiry and an owner who says he did not check it.
Part III also sharpened the question of who was approved to command and who actually did. Cazorla Jr’s owner says the man named as master had never operated his boat. That makes the original certificates, approved conditions and the named men’s accounts essential. It does not, on its own, establish fraud, agency complicity or why the vessels collided.
- Both boats suffered serious consequences.Pascal directly says Cazorla Jr sank after a post-impact steering problem and an attempted move shoreward. Ocean Angel witnesses describe breakup and overturning. Neither is established as an intact, serviceable survivor.
- The warning accounts must stay separate.Elias alleges that women aboard Ocean Angel asked Dylan to slow and he did not; Dylan denies a request. Pascal says Tyson immediately complied aboard Cazorla Jr.
- A precise collision reconstruction remains missing.Recalled times, engine RPM, unseen approaches and a commissioner’s damage proposition cannot establish measured tracks, an impact angle, impairment or fault.
The people behind the inquiry
These are the four names in the police-attributed 22 July account. Later family reporting also identifies Tyrone as Tyrone Paul. That difference needs reconciliation; it does not add another person. Respectful coverage should preserve both identity accuracy and the distinction between evidence about the collision and conclusions about each death.
02 / The vessels
Two craft. Two identity questions. Two damaged boats.
Photographs help identify and understand a vessel. They do not prove its condition on the night, its exact load or the collision mechanics.
Ocean Angel
SLPA’s highly likely working identification is Ocean Angel #3. Converging markings and configuration, together with additional research supplied by Kevin L. Michel, support that assessment. Registry history and image dates are still needed; the number discrepancy cannot safely be dismissed as a typo.
- Journey
- Northbound from Soufrière. Tyrone initially steered; Dylan Nicholas says he held the helm at impact.
- Configuration
- Red-and-white open passenger layout, aft console and one visible outboard in the selected images. Dylan describes one 200 hp engine; the equipment record remains to be checked.
- Aboard
- Nine in the testimony: Dylan, Tyrone, Roxan, Elias, Debbie, Bradley, Marvin, Marley and an unnamed woman from Choiseul. First names and spelling variants are not a certified manifest.
- After impact
- Described as broken, overturned and missing its bow. Survivors describe floating remains and jackets.
Cazorla Jr
The Mango Tours photograph visibly carries SL078180, matching Cazorla Jr in the inquiry notice. This is a strong photographic identifier lead. The photograph’s date and the owner/registry crosswalk still need authentication.
- Journey
- Tyson Hippolit describes an outward trip to Gros Islet and operated the southbound return. Owner Ireanius Pascal was aboard.
- Configuration
- Blue-and-white open passenger craft in the selected photograph. Pascal explicitly confirms 150 hp; that is owner testimony, not a verified engine or speed measurement.
- Aboard
- Nine in the testimony; five linked by name: Pascal, Tyson, Mavado/Janelle Butcher, Taylor River/Rivier and Roman/Rome William. Four remain unidentified in these passages.
- After impact
- Pascal describes disconnected steering, an engine still running, a failed attempt to move towards shore and sinking.
Owner, named master, regular captain and actual operator are different roles.
Matthew’s and Paul’s direct accounts and SLASPA’s detailed explanation are not included in this first-sitting record. Their knowledge, consent, qualifications and the legal effect of the arrangements cannot be inferred solely from Pascal’s account.
03 / People and roles
Who is asking, and who is answering?
The four commissioners, inquiry officers and six oral witnesses have distinct roles. A person listed in the investigation file did not necessarily give evidence that day.
The Commission and its officers
Justice Mario Michel
Chair; retired judge.
Cuthbert Didier
Commissioner; maritime consultant.
Kentry Frederick
Commissioner; retired Police Marine Unit superintendent.
Luke De Freitas
Commissioner; Assistant Commissioner of Police.
Lenita Joseph
Commission secretary; retired permanent secretary.
Shahida Charlemagne-Octave
Legal counsel to the Commission, not a commissioner.
The inquiry may recall witnesses. Commission counsel is not automatically counsel for a witness or family. Its fact-finding and reform work should not be treated as a criminal determination.
The witnesses heard
PART I
Corporal 153 Luciano — surname unverified
Police investigator, Marine Intelligence and Investigations Unit. Tendered the investigation file; described missing documents and outstanding investigative work. Surname needs verification.
PART I
Roxan Bonapart
Passenger on Ocean Angel. Name rendered Roxan Bonapart in the supplied transcript; spelling unverified. Described speed, lights, a life-jacket request and rescue. Did not see the collision itself.
PART I
Dylan Nicholas
Ocean Angel operator at impact; managing director of Ocean Angel Boats. Accepted several omitted precautions; disputed accounts of excessive speed and unlit travel after he took over.
PART II
Elias Nicholas
Ocean Angel passenger; fisherman. Described the operator change, passenger warnings, capsize and rescue. His view forward and memory after injury were limited.
PART II
Tyson Hippolit [spelling provisional]
Operator of Cazorla Jr. Accepted a displayed licence expiry date, admitted beer consumption and described his lights, lookout and rescue actions.
PART III
Ireanius Pascal [spelling provisional]
Owner of Cazorla Jr and aboard that night; tour operator. Gave sworn oral evidence about the substitute operator, named-master paperwork, omitted precautions and his vessel sinking after impact.
Roxan Bonapart is the working form rendered in the supplied transcript; spelling is unverified. Tyson Hippolit and Ireanius Pascal also remain provisional spellings. Variant renderings are preserved in the source record, not treated as additional people.
Explore what each witness said →04 / What happened that night?
Two journeys, an unobserved gap, several rescue accounts.
The accounts establish an order of reported events. They do not supply a common clock or a defensible drawing of the boats coming together.
The coast gives the accounts context.
Ocean Angel’s journey was northbound from Soufrière. Cazorla Jr was returning south after a trip to Gros Islet. The hearing identifies Anse Cochon as the collision locality. Elias describes one rescue landing at Canaries.
Those place names do not establish a vessel’s distance offshore, a route around a landmark, a sinking position or a measured contact point. The map deliberately shows places rather than vessel tracks.
The official 6 p.m. forecast on 18 July warned of locally rough seas within 25 miles and waves or swells of 6–9 ft. This establishes a forecast available before departure, not the wave height or visibility at the collision.
Ocean Angel · Reported northbound journey
A change at the helm
- Departure and takeoverTyrone initially steers. Dylan takes over. Dylan acknowledges an unlit departure, then says he switches lights on.
- Warnings and limited viewsRoxan describes fast travel and a jacket request; Elias reports the women’s slow-down concerns. Both have limited forward views. Dylan disputes the warning account.
- A claimed sighting and turnDylan says he sees red and green lights and alters to starboard. The distance, interval and extent of the turn are not measured.
Cazorla Jr · Reported southbound return
A slowdown and a later stop
- The return from Gros IsletTyson is operating. Pascal’s departure-time and journey-duration estimates are approximate; he did not check the clock.
- A request to ease offPascal says he asks Tyson to slow near Windjammer because of choppy conditions. He says Tyson immediately complies and does not accelerate again.
- A jacket requested for RomanPascal describes a later stop to obtain a jacket, clarifying that Taylor requested it. Wearing at impact and the exact interval are not established.
The central gap
Contact is not reconstructed.
A claimed turn, no reported sighting from the other operator, passengers’ obstructed views and disputed damage propositions do not resolve the impact angle. No collision vectors or individual injury trajectories are asserted here.
Ocean Angel’s aftermath
- Breakup, overturning and flotationDylan describes a broken hull; Elias describes the capsized craft and a floating jacket. Roxan’s memory jumps from the jacket discussion to being in the water.
- Separate rescue recollectionsRoxan describes a boat responding to a phone light. Elias recalls reaching Canaries with Debbie’s body and later losing awareness. Their estimates do not time the whole response.
Cazorla Jr’s aftermath
- Neutral and assistanceTyson says he puts the engine in neutral and enters the water to save two people. Pascal supports his going to help, without verifying the exact two-person count.
- Failed steering, sinking and a callPascal says he and Mavado try to move shoreward with the engine running but steering disconnected. The boat sinks. He reaches a yacht and calls his aunt.
Interactive account explorer
Follow each account at your own pace.
Five witness selections and thirty source-linked scenes. Play, pause or step through the sequence. Reduced-motion and readable text views preserve the same evidence. The visuals explain testimony, not a forensic reconstruction.
05 / Compare the evidence
Some differences matter. Others come from asking different questions.
Read the statements alongside their limits. An admission, an observation, a reported remark and a questioner’s proposition carry different evidential weight.
The table scrolls horizontally on small screens. Select a source reference to read the underlying passage without losing your place.
| Issue | Ocean Angel account | Cazorla Jr account | What the comparison establishes |
|---|---|---|---|
| Speed and warnings | Roxan and Elias describe faster travel after Dylan took over. Elias says three women asked him to slow down and he did not. Dylan denies receiving a request. | Pascal says Tyson immediately slowed near Windjammer when asked and did not accelerate again. | Do not combine the voyages into a claim that both operators ignored warnings. Perceived pace is not measured speed. |
| Navigation lights | Dylan accepts an unlit departure but says he switched the lights on. Roxan and Elias say they observed no lights. | Tyson says his lights were on. Pascal clarifies that navigation lights were in use; the working VHF was the equipment left unused. | Departure, switch position, functioning lamps and what a seated passenger could see are separate propositions. |
| Lookout and approach | Dylan names Tyrone as lookout but says he personally saw red and green lights without a report from him. The passengers did not see the other boat approach. | Tyson says a lookout saw nothing. Pascal names Mavado/Janelle Butcher but retracts certainty about what the lookout could see. | The Cazorla lookout has not testified in this supplied day. A lack of warning does not by itself prove another vessel was unlit. |
| Life jackets | Dylan accepts that nobody aboard was wearing one. Roxan says she asked for one and did not receive it before the collision. | Pascal admits no routine distribution and describes a stop to obtain a jacket for Roman William, requested by Taylor. | Carriage, accessibility, request, retrieval and wearing are different facts. Roman’s wearing at impact is not established. |
| Impact and damage | Dylan describes a claimed starboard turn and a broken, overturned vessel; contact-side questioning exposes unresolved differences. Elias did not see contact. | Pascal describes the impact and post-impact disconnected steering, then an unsuccessful attempt to move shoreward and sinking. | Neither a survival inference nor a damage proposition establishes an angle. Steering failure is described after impact, not established as its prior cause. |
| Survival and rescue | Witnesses describe floating material and jackets. Elias describes rescue to Canaries; Roxan describes a later boat responding to a phone light. | Tyson says he entered the water and saved two people. Pascal supports his going to assist, then describes reaching a yacht and calling his aunt. | The exact two-person rescue count and the complete dispatch, rescue and medical chronology need independent records. |
A closer audit of apparent contradictions
The purpose is to identify the exact proposition that needs testing, rather than treat confusion in an imperfect transcript as proof of dishonesty.
The “15 minutes” is not a proved warning interval.
Dylan repeatedly explains his approximately fifteen-minute estimate as the interval from taking the helm to the collision. Questioning compares it with a much shorter estimated distance to the lights. Those endpoints cannot be substituted for one another. Engine RPM, prompted knot estimates and utility-pole distances also do not make a measured reconstruction.
The consequence for reporting: Establish separate times for departure, helm handover, first sighting and impact before calculating reaction time or closing speed.
Elias’s answer was qualified before the apparent reversal.
A question asks whether anybody “other than the girls” requested a speed change. The chair later restates this broadly as nobody having complained. The final clarification restores Elias’s account that the women complained and Dylan did not slow. The omitted qualifier matters to any assessment of his consistency.
The consequence for reporting: Read the question and answer together. A separate stable/unstable wording problem remains unresolved without the recording; its cause should not simply be attributed to transcription.
Pascal gives several intervals, not one synchronized clock.
His estimates concern the time from slowdown to jacket request, from the request to impact, and from slowdown to impact. The first two intervals can both be true. His separate fifty-to-sixty-minute estimate creates the need to reconcile all three endpoints.
The consequence for reporting: Retain the request and stop as testimony, while leaving their exact place in the sequence open. Do not infer Roman wore the jacket at impact or that it determined survival.
An expired licence and a later expiry are different issues.
Tyson accepts the displayed 18 July 2015 expiry. The hearing separately discusses Graham Matthew’s 28 September 2026 expiry and a vessel-certificate expiry in November. September follows the July collision; different documents can also have different expiry dates.
The consequence for reporting: Report Tyson’s admission accurately and seek the original record and renewal history. Do not backdate Matthew’s September expiry or treat nonmatching document dates as an offence by themselves.
A discrepancy in staffing warrants documents, not a fraud verdict.
Pascal says Graham Matthew was named as master but never operated his boat, Daryl Paul was its regular captain and Tyson operated that day. He says the insurer accepted Paul while SLASPA sought a different licence level. The legal conditions, approved substitutions and the absent named men’s accounts are not supplied here.
The consequence for reporting: The admission establishes a serious paper/practice question. It does not establish Matthew’s consent, Paul’s lack of all qualifications, insurance invalidity, forgery or official complicity.
Consumption, demeanour and survival are not technical measurements.
Tyson admits at least four beers, but the record does not establish his blood-alcohol level or impairment. Pascal’s tone is challenged; that does not establish indifference. The suggestion that a forward lookout should have been killed or badly injured is also not a reconstruction.
The consequence for reporting: Use direct observations, contemporaneous records and appropriate technical evidence. Neither a panicked first call nor a person’s survival proves the cause of the collision or an avoidable individual death.
187 source-located entries
Inspect the claim, its status and its limit.
The evidence explorer holds the full register; thirty-seven editorial selections provide a starting point. Search by person, vessel or issue, and move from an analytical statement to its exact transcript location.
06 / Editorial assessment
A productive sitting. A test of the questioning, too.
The Commission exposed serious operational questions and widened the record. Its next task is to connect pointed questioning to authenticated documents, direct witnesses and explicit reasoning.
Michel set out the remit, explained representation and safeguards, and kept open the possibility of recall. The panel’s strongest exchanges narrowed broad assertions into checkable claims. Its weaker exchanges risked letting an interpretation become the premise of the next question before the underlying fact had been established.
That is a reason to examine the process carefully, not to dismiss the inquiry. A public record becomes stronger when its corrections and limits are visible. The same standard should apply to this coverage.
What strengthened the record
Procedural care and useful clarification
Representation and participation received concrete attention.
The hearing allowed legal consultation, including a private lawyer–client discussion, and accommodated Dylan’s medical condition and Pascal’s hearing difficulty. Those arrangements matter when hesitation or repetition might otherwise be mistaken for evasion.
The file was treated as a record to complete and test.
Missing Pascal statement pages were recovered from the maritime report. The investigator’s principal role that morning was distinguished from his future substantive evidence, and further scene material and logs were requested.
Some questioning narrowed evidence usefully.
The name-by-name discussion produced a working nine-person Ocean Angel list. Pascal withdrew his certainty about what the lookout could see, and the limit was accepted. Systemic questions about reporting, licences and briefings also went beyond the instant of impact.
Where the next phase needs precision
Questions should not outrun proof
Keep a witness’s qualification in the question.
Dropping “other than the girls” manufactured an apparent conflict in Elias’s answers before the chair clarified the position. The correction is valuable; the intervening restatement should not be used as proof of unreliability.
Do not let technical authority outrun the inputs.
Prompted distances, engine RPM and recalled intervals do not independently establish speed or reaction time. A claim to have switched on lights and a passenger’s failure to see them require positional and equipment evidence before a binary credibility verdict.
Keep legal and causal propositions open to proof.
The staffing discrepancy is important, but a question linking it to four deaths remains a question. A suggested injury trajectory or an impression of casualness cannot substitute for physical evidence or establish a person’s state of mind.
07 / The evidence still needed
Fourteen questions that should produce records.
These are evidence requests and explanatory questions, not allegations disguised as questions. The status is tied to the 13 October record and the materials reviewed here.
The police file was incomplete at the sitting; Pascal’s missing pages were found during the hearing in the maritime report. A record described as outstanding then may since have been obtained. We do not infer concealment or that an agency never held it.
01What was each operator authorized to do on 18 July?
02Why did the documented and working captains differ?
03Who was on each craft, in which role and position?
Records and explanation: Booking/departure records, contemporaneous lists and surviving occupants’ evidence. A witness-checked seating plan would inform sightlines without guessing where people sat.
04Which lights functioned and were visible from which positions?
Records and explanation: Lamps, wiring, switches, maintenance and inspection records, photographs and examinations of the physical remains. Ask each witness about their own viewing position.
05What is the defensible chronology?
Records and explanation: Phone metadata, VHF and emergency logs, departure evidence and any navigation/engine data. Keep estimated intervals and known clock times separate.
06What do both hulls establish about contact and movement?
Records and explanation: Scaled surveys, original damage images, recovery positions, engine/steering examination and a reconstruction that states its uncertainty. Include registry histories for both Ocean Angel numbers.
07What did the lookouts actually do and perceive?
Records and explanation: Direct evidence from the Cazorla lookout, relevant Ocean Angel crew statements and records of positions and communication. Distinguish assigned role from demonstrated action.
08How did distress information reach responders?
Records and explanation: Call, dispatch, Marine Unit and rescue-vessel logs; the relevant phone records and direct accounts of civilian rescuers. Reconstruct the chain through medical care.
09Which safeguards applied, and what was actually practiced?
Records and explanation: Passenger limits, certificate conditions, safety-briefing and jacket requirements, equipment inventories and operating policies. Clarify any claimed difference between tourist and local outings.
10What evidence bears on alcohol, fatigue and fitness?
Records and explanation: Relevant contemporaneous medical evidence, observations, drink timing and work/rest history. Recalled quantities and experience cannot alone measure impairment at impact.
11What caused each death, and which survival opportunities existed?
Records and explanation: Appropriate pathology evidence, recovery/medical chronology and expert survival analysis, handled with family sensitivity. Avoid assuming a particular precaution would certainly have saved an individual.
12Where did oversight operate or fail, if anywhere?
Records and explanation: Inspection, boarding, patrol, renewal and enforcement records, lighthouse procedures and staffing. Witness memories of being checked are not a complete institutional audit.
13How will the evidential record be completed and made accessible?
Records and explanation: A dated exhibit index, outstanding-records tracker, corrections, recall timetable and clear arrangements to answer adverse propositions. Distinguish hearing access from access to exhibits and transcripts.
14Which interim safety measures follow from evidence already secured?
Records and explanation: A reasoned response with responsible offices, resources, dates and review measures. Match action to demonstrated risks without awaiting or pre-empting every causation finding.
The supplied record does not include SLASPA’s detailed explanation of the certification decisions or direct testimony from Matthew and Paul. Their accounts and the original records could materially change this analysis. This is not a claim that any person or institution refused comment.
08 / From testimony to safer voyages
What does SLASPA have to do with this?
The Division of Maritime Affairs is relevant to registration, survey and certification, operator competence, manning requirements, casualty investigation and the interface with rescue. A remit is not itself proof of a failure.
The practical question is what each safeguard was supposed to do before these voyages. A certificate should connect a vessel and its operating conditions to competent, authorized people. An inspection should identify deficiencies and record their closure. A departure process should leave information usable in an emergency. A rescue system should preserve the chronology through dispatch, recovery and care.
Where testimony suggests a safeguard was absent, the inquiry should establish the rule, the responsible person or office, the actual check and the result. That separates an omitted precaution from proof that it caused the collision. It also permits carefully justified prevention before every responsibility dispute is resolved.
Map the existing duties.
Connect the actual vessel, service and operator to the applicable certificate, licence and safety requirements. A new law should not obscure a control that already existed.
Find the record of the check.
Examine staffing, briefings, jackets, lights, departure information, inspections and enforcement. Distinguish a rule on paper from a safeguard operating in practice.
Specify the change and its owner.
Where evidence supports a defect, define the remedy, lawful authority, responsible office, resources and review date. Use independent technical scrutiny where needed.
Track safer operation.
Publish inspection coverage, serious defects closed, response readiness and implementation progress with useful denominators. More reporting can initially mean better detection.
The Maritime Act connection
Evidence should test the reform proposal.
The inquiry’s remit includes recommendations to inform modern maritime legislation. SLPA’s proposed Maritime Act and implementation programme remain proposals, not adopted law or Commission recommendations. The first sitting offers specific tests for the design.
Keep the public inquiry, the separate statutory safety investigation and criminal/civil processes distinct. They have different purposes, legal powers and evidence protections. Read the existing-law analysis and reform framework.
09 / Sources, method and corrections
The analysis should be as checkable as the claims.
Read the transcript text, inspect the evidence register and review the photographic provenance. The public-facing documents preserve attribution and the specific limits of the available record.
What this edition uses
- Three supplied automated transcripts covering the first sitting. Part I locators are recording offsets; P2 and P3 references are editorial passage IDs, not clock times. They are not certified verbatim records.
- Official appointment and sitting notices, used for the Commission’s roster, remit and what each notice prints. A public notice does not authenticate an exhibit discussed in testimony.
- Police-attributed and family reporting for named deaths and dated public developments, distinguished from direct Commission evidence.
- Selected source photographs, with dates and provenance limits retained. Different encodings of the same image are not independent corroboration. No pictured person is identified from appearance.
- SLPA editorial analysis, including additional research supplied by Kevin L. Michel, an evidence audit, source-linked questions and explanatory visualizations. Photographic identity confidence does not determine fault.
Statements are paraphrased unless quotation marks appear. Quoted wording reproduces the supplied text and has not all been checked against the recording. Original attachments, organized passages and analytical claims remain different records. The animated accounts depict sequences of testimony, without measured tracks, timing, impact angles or individual death trajectories.
The chair announced a next sitting for 14 October at 10 a.m. This first-sitting record does not establish that it took place or include subsequent evidence.
Explore or download the record
Principal public sources
Corrections and distinctions incorporated in this edition
· The following points are explicit to prevent misleading reuse of the record.
- Registry numbers: the notices print different sequences, SL077875 and SL077857. The discrepancy is unresolved; no registration change or publication error is established.
- Elias’s warnings: the “other than the girls” qualification is retained; the later broad restatement is not treated as an independent witness reversal.
- Jacket timing: slowdown-to-request and request-to-impact are separate intervals; the additional slowdown-to-impact estimate needs reconciliation. Wearing at impact is not established.
- Dates and names: Matthew’s September expiry is not a July expiry. Provisional name forms are labelled, and people named in a file are distinguished from the six oral witnesses.
- Visual identification: Ocean Angel #1 and #2 remain outside the active candidate group. The selected likely-#3 photographs and the generated layout models have separate evidential roles.
Earlier coverage and the continuing inquiry
The research history, appointment documents, submission information and earlier reform framework remain available in the policy context. The 9 October sitting announcement is preserved as dated history, alongside the completed first-sitting account. Public attendance, access to exhibits, transcript access and release of a final report are separate questions.