Surviving the Tarawera eruption: the experience of Māori settlements in the Red Zone

A. Rex Bunn 19 Surviving the Tarawera eruption: the experience of Māori settlements in the Red Zone A. Rex Bunn The relations of large objects to one another, can be observed better at a distance than close at hand, when we see nothing clearly except what is immediately next to us. James Froude, Oceana 1886. 1. Introduction The 1886 Tarawera eruption was the first major volcanic eruption seen in colonial New Zealand. It was unique in its 4·5-hour eruption from multiple vents along a mountain and sub-lake fissure of 15–18 km. Much investigation has been made of the geology and volcanology (Nairn 1979, Rowe et al). Less has been made of the casualties and even less of the casualties’ locations, outside the major village of Te Wairoa. Differing estimates of mortality are 90–150, with ~ 120 being the recent estimate (Lowe 2001). Virtually no investigation has been made of lake residents who may have survived one, two or all phases of the Tarawera-Rotomahana eruptions that June night. The timeline of the eruptions varies by observer and location. This is due to several factors. It began late at night and observers awoke at different times with the ground shaking. There was no town clock, nor likely a chronometer at Te Wairoa. Uncalibrated watches recorded different times. The shaking intensity varied according to the observers’ distance: those closer awakened earlier. No official report author witnessed the eruption— hence the government investigators relied on reports from folk they interviewed on brief visits to the region. The interviewer and interviewees were not always bilingual. There was no inquest. There was a racial bias with colonists providing nearly all the interviewees. Few Māori reports were published in English, other than those of Alfred Warbrick (1860–1940), Sophia Hinerangi (c.1834–1911), Kimihia (fl. 1886) and Willy Bennett (1874–1959). Māori reports were often untimed. This is a pity, for those nearest the eruptions, and those who survived from close quarters, were Māori. Alfred Warbrick was probably the closest observer (10 km) and gave one of the best-timed eyewitness descriptions yet was largely ignored by later researchers. Huta Tangihia (fl. 1886) was even closer (5–6 km) but wisely did not stay to witness the eruptions. Based upon eyewitnesses and with a modern understanding of the eruption, I assemble a fresh timeline drawing on Māori and colonial sources with a priority for eyewitnesses who were close to at least one stage of the eruptions, with a watch. Establishing the headcount of those around the eruption Red Zone (6–10 km from the Rotomahana vents) on the night is difficult, owing to poor records and census data for Māori. There were ~84 Māori settlements and gardens around the Tarawera lake and massif. Oddly, they are not catalogued. Some are undocumented, and locations are confused. There was uncertainty over which settlements were occupied. It was winter on the plateau and the author attests to cold winters that see snow on Mount Tarawera. Outdoor activities, such as gardening, would be at a low ebb, although the season was late that year. The region had not been surveyed by the colonial government before the eruption. Fortunately, the settlements had been mapped by Ferdinand Hochstetter (1831–1884) when AJVS 29: 2 (2025) General Issue 20 he surveyed the central North Island in 1859. This was part of the German contribution to New Zealand science— small in number but profound in impact. Ernest Dieffenbach (1811– 1855), Hochstetter (aka the Father of New Zealand geology), and Julius Haast (1822–1887) essentially wrote the early geological and natural history of New Zealand. 2. Methods 2.1. Location fixing The first step in seeking survivors is to fix their likely locations at 11.30 pm on Wednesday, 9 June 1886. The first pre-eruption, large-scale mapping of the lake settlements around Lakes Tarawera, Rotomahana, Rotomakariri and Rerewhakaaitu was made by Ferdinand Hochstetter as part of his geographic and geological maps and survey (see Figure 1). These were digitally repatriated to New Zealand by Sascha Nolden between 2010 and 2015. Hochstetter’s central North Island survey was reconstructed and published by Nolden and the author (Bunn and Nolden 2016, 2018, 2023). Fig. 1. Hochstetter’s survey map of Lakes Tarawera, Rotomahana, Rotomakariri and Rerewhakaaitu (From: S. Nolden & S. B. Nolden, Hochstetter Collection Basel: Part 3 – New Zealand maps and sketches, Auckland: Mente Corde Manu, HCB 3.5.6 (detail)). The settlements were elaborated by searches of the New Zealand Archaeological Association (NZAA) ArchSite, news media and reports from the Rotorua District Council, the Tuhourangi Tribal Authority, Ngati Rangitihi Hapu, Heritage New Zealand and Te Arawa Lakes Trust. The latter report draws on NZAA records. In turn, these rely on 1960s research by Cecil Watt (fl. 1961) and Ken Moore (fl. 1971), with recent additions by Ken Phillips. Ron Keam (1932–2019) assembled an unmatched eruption archive, but confined himself to description, rather than a forensic analysis (Keam 1961–1978, 1988). A. Rex Bunn 21 To begin, Hochstetter’s 1859 Observation Station 20 on Kakerangi was located on the high contour above Oneroa. His large-scale map shows the Tuhourangi track linking Te Wairoa with Lake Rotomahana. In Figure 2, bearings from Station 20 to seven landmarks around Lake Tarawera were transcribed, translated and declination-adjusted using the National Oceanic and Atmospheric Administration (NOAA) Historical Declination Viewer https://www.ncei.noaa.gov/maps/historical-declination/. Fig. 2. Hochstetter’s bearings from his Observation Station 20 on Kakerangi (green rays). The umber rays are canoe escape courses from the Ngutuahi Arm to Wairoa (Bunn/Google Earth™). The Station 20 location was validated by a seven-bearing resection using reciprocal bearings from Rua-O-Mukeria (aka Kariri), Otangimoana, Otamatea, Oneroa, Moura Point, Pukekiore Point and Ohume Landing. The Kakerangi bearings enabled further sites to be triangulated from adjacent bearings, e.g. Otamatea and Oneroa lay on the same bearing. The Tikauwe (aka Te Kauae) island pa was located on the shoal off Oneroa. Wheturau was located after correcting the NZAA Otamatea location with a survey bearing. Oneroa was connected with the lost Totarariki and Matakana sites using combined Hochstetter and NZAA data. This assisted the review of eruption survivors and fatalities. An 1841 Dieffenbach bearing on Lakes Rerewhakaaitu and Rotomakariri assisted with the location of Te Ariki and Otuapane. 2.2. Tarawera settlements and gardens Using the combined resources, a database of 46 settlements around the TaraweraRotomahana Basin is presented in Table 1. Many were unoccupied on June 9. AJVS 29: 2 (2025) General Issue 22 Table 1. The 46 sites around Lakes Tarawera and Rotomahana (clockwise from six o’clock.) In Table 2, Watt and Moore documented 41 sites as gardens, linked to the settlements. Settlements around Lakes Tarawera and Rotomahana, as at June 1886. Names Location Type NZAA Code Affiliation Latitude S Longitude E Waingnongogno Lake Rotomakariri Kainga n/a Ngati Rangitihi ~38.2560 176.4471 Puai Island Lake Rotomahana Kainga n/a Tuhourangi ~38.2632 176.4302 Pukura Island Lake Rotomahana Kainga n/a Tuhourangi ~38.2633 176.4323 Patiti Island Lake Rotomahana Pa V16/2 ~38.2680 176.4453 Te Ariki-Piripai1 Rapatu Bay Kainga n/a Tuhourangi ~38.2466 176.4519 Kokotaia Lake Tarawera Pa V16/70 Tuhourangi ~38.2466 176.4523 Ngawhiro Lake Tarawera Kainga V16/264 ~38.2460 176.4531 Pukekiore Lake Tarawera Pa V16/72 Tuhourangi ~38.2458 176.4425 Piripai2 Te Rata Bay Kainga n/a ~38.2455 176.4408 Oneroa Oneroa Bay Kainga U16/28 Ngati Apiti & Ngati Taoi ~38.2266 176.4144 Te Kauae Lake Tarawera Pa U16/30 Tuhourangi ~38.2256 176.4148 Otuapane Lake Tarawera Kainga/Garden U16/26 ~38.2425 176.4165 Hakaipari Lake Tarawera Pa U16/27 Ngati Putu, Taoi &/or Apiti ~38.2394 176.4209 Totoariki Oneroa Bay Garden n/a Tuhourangi 38.2269 176.4129 Mataru Makino Lake Tarawera Pa V16/17 Ngati Taoi ~38.2303 176.4389 Moura Lake Tarawera Pa V16/18 Ngati Rangitihi ~38.2177 176.4397 Tawhitinui Lake Tarawera Pa U16/29 ~38.2196 176.434 Hawaiki Lake Tarawera Kainga U16/146 ~38.2133 176.3974 Kakanui Lake Tarawera Pa U16/31 ~38.2129 176.3954 Karikaria Lake Tarawera Pa U16/4 Ngati Hinemihi ~38.2077 176.3805 Te Wairoa Lake Tarawera Town U16/23. Tuhourangi ~38.2125 176.3631 Punaromia Lake Tarawera Rock Art U16/11 ~38.2040 176.3785 Te Kawau Lake Tarawera Kainga U16/32 Ngati Hinemihi ~38.2040 176.3785 Ruaomukeria Lake Tarawera Kainga n/a Ngati Hinemihi ~38.2028 176.3813 Kariri Lake Tarawera Pa U16/15 ~38.1994 176.3956 Waitangi Lake Tarawera Kainga U16/60 ~38.1792 176.3886 Waitangi Lake Tarawera Stream U16/187 ~38.1781 176.379 Tokoniho Lake Tarawera Kainga n/a Tuhourangi ~38.1976 176.3719 Te Karamea Bay Lake Tarawera n/a U16/61 ~38.1734 176.3922 Otumutu Lake Tarawera Pa U16/8 ~38.1684 176.4003 Whareroa Lake Tarawera Pa V16/258 Tuhourangi ~38.1692 176.3973 Otamatea Lake Tarawera Pa V16/73 ~38.1809 176.4719 Wheturau Lake Tarawera Kainga n/a ~38.1767 176.4527 Tapahoro Lake Tarawera Pa V16/58 Ngati Rangitihi ~38.1842 176.5015 Kanaehapa Lake Tarawera Pa V16/288 ~38.2071 176.4584 Te Puna Lake Tarawera Kainga V16/287 ~38.2243 176.4502 Te Koutu Lake Tarawera Pa V16/69 Tuhourangi ~38.2366 176.4553 Otangimoana Lake Tarawera Kainga n/a ~38.1658 176.4299 Ohune Lake Tarawera Landing n/a ~38.2124 176.4508 Makatiti Lake Tarawera Hut n/a ~38.1409 176.4663 Ruawahia Tarawera massif Garden n/a ~38.2575 176.5577 Pakaraka Tuminui Kainga U16/93 Tuhourangi ~38.2491 176.3191 Manganui Lake Tarawera Kainga N86/6 ? Onepoto Lake Tarawera Kainga N86/6 ? Matakana Oneroa Bay Garden n/a Tuhourangi -38.2273 176.4147 Pahipoto Rangitaiki River Kainga n/a Ngati Rangitihi -38.0103 176.8137 A. Rex Bunn 23 Table 2. 41 Tarawera Garden Locations from NZAA records. These contain the first locations for Totarariki and Matakana at Oneroa Bay. The above 87-site database compares with the 29 NZAA site records and the ≤ 100 Tarawera sites recalled by tribal historians. 1 2.3. Māori mortality clusters Next, the locations of major casualty events were ranked. This disclosed errors in the historical record, which reports that the major Māori eruption casualties occurred at Te Ariki (27–52 deaths), with the second cluster at Moura (25–39 deaths). A third, unrecorded mortality cluster of 23 deaths is herein reported at Oneroa Bay. A fourth of 18 deaths is now likely at Waitangi. The Te Ariki mortality presented a problem. It is the most frequently mentioned site in the historiography, but this research shows there were multiple locations termed Te Ariki. 2.4. Māori demographics The Māori demographics are poorly documented, as New Zealand censuses at the time did not gather Māori demographic data. The numbers included half-castes living as Maori [sic]. In the Rotorua Riding, there were 1,377 Māori recorded in the May 1886 report (Brabant 1886). The 1881–1886 decline was estimated at ~ 1·25%. The census noted respiratory diseases as the leading cause of death for the young and old. 2.5. The Red Zone Given that most eruption mortality was assumed due to base surges from the RotomahanaOkaro vents, a Red Zone was established around the Rotomahana vents, using surge evidence from Ian Nairn in 1979 (Nairn 1979). Nairn sketched a circular 6 km surge range and published their direction arrows in Figure 3. In 2001, Lowe and Keam reviewed the eruption mortality and extended the Red Zone to 10 km from the Great Crater at Rotomahana in Figure 3 (Lowe 2001). This encloses the area experiencing the impact from surges, ballistic 1 For sites with Hochstetter’s bearings and a map location on/near the shore, the coordinates should be precise to four decimal places at this latitude, i.e. the latitude precision is ~ ± 5·6 m and for longitude ~ ± 4·4 m. For those sites with a map location, I suggest a precision of three decimal places ± 55·7 m latitude and ± 43·9 m longitude. This is adequate for imaging search-boxes. AJVS 29: 2 (2025) General Issue 24 rocks and stones, mud sky-fall, winds and ash, affecting respiration and movement. Outside this, the impact was presumed to lessen. Fig. 3. Map of the Rotomahana Basin c. 2001 showing base surge directions and ranges. (Courtesy David Lowe). The blue images for Lakes Rotomakariri and Rotomahana predate this author’s 2014–2025 research, setting their more northerly location. In Figure 3, the 10 km radius includes Moura and Oneroa Bay. 2.6. Tamihana Te Keu, Matakana and Totarariki at Oneroa Bay After plotting the 46 Tarawera settlements, there is a suspicious lack of sites along Oneroa and Te Rata Bays. These have prime, north-facing land for homes, fishing and gardens. From photography, we know that, for example Piripai had attached gardens. Watt, Hochstetter and Percy Smith (1840–1922), plus photography of Te Ariki, suggest the answer may lie with Tamihana Te Keu, the Chief of Rotomahana. The published prints of Te Ariki (aka Piripai, aka Philippi) are of the same settlement and garden. Te Keu and family members are prominent in prints. Watt had access to Māori sources. Some of his records are held in Whakatane. Te Heu, his wife and seven children died in the eruption at Totarariki (Star June 14, Keam 1988). The locations of Totarariki and Matakana were undocumented (Keam 1988). There were no inquests establishing the death toll from the 1886 eruption (Brabant 1887). In his first compilation of Māori casualties, the 14 Matakana deaths were assumed to be doublecounted (Keam 1988). Yet, in 1961 Watt had identified Totarariki and Matakana as gardens A. Rex Bunn 25 vicinal with Oneroa and its island pa Te Kauae. This is important, for these deaths form the third largest mortality cluster (Herald, June 16). 2.7. Logistic regression analysis This is a statistical method, measuring the factors affecting an outcome. In this case, we wish to find the factors helping or hindering the Māori residents in surviving the Tarawera eruption. In a regression model, we plug in risk factors to assess their importance, while holding others aside. The method is now used in many disciplines. In the 1970s, the author pioneered such path regression with beta coefficients in health economics research (Bunn 1978, 1979a, 1979b, 1980). This is the first application to the Tarawera eruption. For this research, 13 occupied settlements (kainga) were selected, i.e. Wairoa, Puai and Pukura Islands, Rapatu Bay, Te Rata Bay, Oneroa Bay, Moura, Makatiti, Ruawahia (2), Tokiniho (aka Tokoniho), Waingongongo, Waitangi and Pahipoto. In some, the residents died, and in others, they escaped. The dependent variable is Survival, i.e., whether the Flight or Freeze decisions made by residents were correct for their survival. The independent variables included the distance from Tarawera and Rotomahana vents, mobility, health, reaction-time, visibility, leadership, objective, motivation, path access, path quality, group size, flight or freeze, travel choice and eruption phase. Model iterations were made with small groups of factors to identify the important variables. This first application was constrained by the small number of occupied settlements and the demographic data. The historical records emphasise European mortality and survival. Lake settlements with known occupancy and outcome are a subset of the total, so we cannot generalise. The subset is formed by three factors. First, most pa were unoccupied as the region was free of conflict. Second, there had been a disease outbreak in Wairoa in AprilMay, with the funerals attracting mourners from around the lake. Thirdly, it was winter, and outside activities were at a seasonal low. Also, many settlements have not been surveyed or excavated. For Te Ariki, confusion persists over its site. The computation of logistic regression is now commonplace with online calculators. Further statistical details are available from the author. 2.8. Time and Distance modelling of eruption escape paths Humans facing a threat trigger the fight-or-flight mechanism, the evolutionary sympathetic nervous system response to a stressor. After recognising a stressor like the Tarawera eruption, the Māori and Pakeha reaction would be Flight or Freeze, the latter in the hope that the eruption would pass by. Those electing a flight strategy take a direction away from the threat. Rotorua residents fleeing on foot in pyjamas took the west road. Given the topography with few paths, those Māori fleeing the lake settlements could choose between trails or a canoe. Employing Hochstetter’s survey and maps, together with historical trail maps and the NZAA archive, we can, for the first time, fix coordinates for the settlements known to be occupied, i.e. Puai Island, Waingongongo, Moura and Oneroa Bay. We can also trace the paths residents would take. Given the regression constraints, the key variables of distance from the vents and access to an escape path were modelled using a Time and Distance model. The settlements included Puai Is. (fleeing to the North West, South and Southeast), Waingongongo (North, West and South), Rapatu Bay (North, West, South and canoe), Te Rata Bay, Otuapane, Moura (by canoe) and Oneroa. AJVS 29: 2 (2025) General Issue 26 2.9 Causes of death and the gas cloud No published forensic analysis could be found of the causes of death for Māori around Lake Tarawera. Given that most bodies were not recovered, there is a lack of post-mortem data. Roof collapses, asphyxiation (from ash particles), and burial in falling mud and stones are mentioned. There are reports of distressing sulphurous smells in Te Wairoa and Rotorua, but these were not interpreted as a hazard. Rotorua residents and tourists quickly undergo olfactory desensitisation (nasal fatigue). That night, this was a hazard as at concentrations >100 ppm, H₂S desensitises the olfactory nerves, if the level rises to a lethal concentration (Bunn 2023c). H₂S > 300 ppm causes respiratory distress and collapse, while at 700– 1000+ ppm, immediate unconsciousness and death in minutes (NIOSH 2007). 3.0. Findings The first survey mapping of Māori settlements around Lakes Tarawera, Rotomahana and Rotomakariri was by Hochstetter in 1859. Māori chiefs gave him the names of landscape features while spurning local surveyors. They learnt that disclosure led local surveyors to sequestrate their land. Hochstetter had their confidence as he took their confidential information overseas. 3.1. Lake Tarawera Hochstetter described the lake from Te Mu: “… Its general form, excluding of its deep side coves, is that of a rhombus, with its main diagonal running from West to East.” (Hochstetter 1867, 405). This remains true, and from the Landing, the eye is drawn to the eastern end at Tapahoro and the southern shore along Oneroa Bay. By boat, the southern Ngutuahi arm is invisible until one closes with Moura Point and its shoals. From the Landing, the pre-eruption lake could have its head at Oneroa Bay, with the Ngutuahi Arm a river discharging into it at Moura Point. Hochstetter saw the Ngutuahi Arm as an extension of the Kaiwaka Channel to Moura. 3.2. Finding Te Ariki After Te Wairoa, the most frequently mentioned place around Lake Tarawera in the nineteenth century tourist literature and photography is one called Te Ariki or Piripai. It was a waypoint for tourists en route to the Pink and White Terraces. Before that, Te Ariki was a term for the Ngutuahi Arm, hence any settlement in the arm could be labelled Te Ariki. There are many locations for Te Ariki in the literature. As the place with the greatest eruption mortality, it’s vital we locate it. The research to establish the location is included in the appendix. This led to undocumented Māori centres and mortality in Te Rata and Oneroa Bays. It seems likely that the term Te Ariki (for the Ngutuahi Arm) became a collective term for Kokotaia pa, Piripai and Ngawhiro in Rapatu Bay, and probably the Piripai in Te Rata Bay. It is possible that Oneroa and Te Kauae were included. 3.3. Totarariki, Matakana and Tamihana Te Keu at Oneroa Bay This analysis is the first to triangulate Oneroa and Te Kauae, using Hochstetter’s survey and Watt’s notes. These establish the satellite gardens of Totarariki, Te Ture O Uenuku, A. Rex Bunn 27 Matakana, and Te Waitoitoi along Oneroa Bay. This helps locate where Chief Tamihana (Thompson) Te Keu and his family died at Totarariki. Watt reports Totarariki was a cultivation near Oneroa. This north-facing ridge offered prime gardening with stream 1050543 for watering. Oneroa Bay offers easy landing, and the bathymetry shows it was in pre-eruption days. Te Keu’s home lay at Oneroa Bay. On June 13 1886, Joseph Warbrick (1862–1903) passed Te Keu’s buried house en route to Te Ariki, five miles from Wairoa and four miles from Te Ariki (5/9 = 55.6 %), (Star June 14, Press June 15). Te Keu’s wife and seven children were being excavated by relatives. Keam reports Te Keu dying there with his wife and family, hence the Oneroa death toll is nine. The Tuhourangi Trail to Rotomahana was mapped by Hochstetter and passes Oneroa. As the crow flies, Oneroa is 55% of the way. Following the trail, it is 56%. These figures show Te Keu did not live at Te Ariki in Rapatu Bay or Te Rata Bay. This finding suggests the photograph of Te Keu and children in Figure A5 was exposed at Oneroa, and the adjacent cultivation is Totarariki. Tamihana (Thompson) Te Keu was known by colonists as The Thundercloud of Rotomahana and levied tolls on tourists (Williams). He represents a bridge to the tourism economy of Wairoa, which replaced the agrarian economy. In the Te Ariki arm, although tourism supplied fees and food income, gardening remained active at Moura, Oneroa Bay, and probably at Te Rata Bay. The four garden sites along Oneroa Bay help answer a question that puzzled researchers. Photography showed too few homes for a large population. It was spread across Rapatu, Te Rata and Oneroa Bays. A larger population (with tourists) in the Ngutuahi Arm required more gardens to balance the koura and inanga from this best Tarawera fishery. The 23 residents who died in Oneroa Bay comprise the third largest group of fatalities after Rotomahana and Moura. Keam assumed the Matakana deaths were double-counted (Keam 1988). They must now be added to the toll and their identities sought. Keam seemed unaware of Watt, Moore and their NZAA records, which date from 1961. 3.4. Māori demographics, health trends and morbidity As a health economist-diagnostic radiographer, the author reviewed the risk factors predisposing the Māori residents to a greater health impact. During the previous century, the indigenous population halved from introduced diseases, conflict, poor housing and health care. Respiratory diseases were the common cause of death amongst Māori children and the elderly. This decline helps explain the unoccupied settlements and frequent funerals at Te Wairoa in 1886. It is likely that Te Wairoa mortality was above the national average, due to tourism. The tourists imported infectious viral and bacterial diseases from Europe and ports along the sea route to New Zealand. STIs would be imported by the young, affluent, male tourists to Te Wairoa. Māori lake settlements, having less intimate contact with tourists, would suffer less. Tuberculosis, pertussis, flu, bronchitis and pneumonitis predisposed Māori residents to acute respiratory distress when inhaling irritant volcanic ash. The 1886 Māori census reported a month before that ... the deaths being generally those of children and old people. The adults who die early ... suffer from lung complaints. Low fever has been known to exist from time AJVS 29: 2 (2025) General Issue 34 before the eruption, Whatapoho fumarole near Puai had auto-ignited after reaching combustible concentrations of H₂S and/or hydrogen (Massy 21, Bunn 2023c). A gas cloud including CO₂ and H₂S was emitted before and during the eruption. In Te Wairoa, there were reports of pungent gas— the trademark Rotorua H₂S (rotten egg gas). The geography explains what happened at Moura. The gas cloud was emitted by the shaking and volcanic action south of the Ngutuahi Arm. It blew into it, settled over the water and wafted north into the Moura Narrows. Here, the arm narrowed by 79%, forming a venturi. With friction, a boundary layer formed on the shores, slowing the gas cloud and creating a pocket of toxic gas. This killed the residents. It also intercepted any evacuees from Rapatu and Te Rata Bays and Otuapane. The cloud blew towards the north shore, Tokiniho, Waitangi and Te Wairoa, where it was detected but no longer lethal, three days later when the first boat team had such respiratory distress, that they fashioned masks from their clothes (Warbrick 1934 68). This cloud explains why the survivors were elevated, i.e. at Makatiti and Ruawahia, while the fatalities were nearly all in the south, at lower kainga, i.e. Moura, Piripai, Oneroa and Puai. While taller animals survived, dead rats were everywhere on the surface, a “canary in the coal mine” (Star June 14). 3.13. The case of Moura village The Moura evidence differs from other sites, i.e. Wairoa, Tokiniho, Oneroa and probably Waitangi, where residents sheltered. At Moura, the forensic evidence suggests they assembled outside. No pre-eruption photographs at Moura are known. Figure 5 shows Moura Point, looking north with Moura village buried. Froude describes Moura as “a Māori farm on the water’s edge. There were boats and nets hung up to dry, a maizefield, an orchard and a cabin.” (Froude 1886). The < 45 residents needed 10–20 homes. A. Rex Bunn 35 Fig. 5. The site of the destroyed village of Moura, Tarawera, circa 1887, Dunedin, by Burton Brothers, Frank Coxhead. Te Papa Tongarewa (LS.004507). Taken shortly after the eruption, as the lake rose. From 1903, bodies were exposed, consistent with escapees being overtaken by toxic gas in the open (NZ Times, Keam 1988). Multiple exposures occurred over 1904–1905, when the lake level fell as the eruption dam on the Tarawera River breached (Herald, 14 November 1904). The Irish travel writer Beatrice Grimshaw (1870–1953) noted in 1905: “On the shores of the lake ... at one spot, the whitened bones of some of the natives who perished ... The name and titles of one who was a great chief are painted on a rock that overhangs the shore.” This would be on the south hillock, with the insignia of Chief Hakaraia Peraniko. The lake fell ~ 4 m after November 1, 1904 from its peak of ~ 303 m a.s.l. (Bell, Hodgson and Nairn). The skeletons were exposed as the lake fell to 298–299 m. This is 7–8 m above its preeruption level and indicates the fatalities occurred close to the shoreline (given house sites allowed for lake rise and fall). Gilbert Mair reported the houses were washed into the lake by a mudslide, with three feet of mud cover (Star June 17). Warbrick says they were simply buried in mud to 75 feet. Figure 5 has no evidence of a mudslide and the landform indicates a shallow mud cover. Fitzgerald, who scuba-dived at Moura, found tree trunks embedded in the lake floor (cf. the sunken totara forest in Lake Rotomahana), but no houses (Fitzgerald, Bunn 2020). On balance, the residents did not die in their homes or in the lake. I suggest they assembled outside and were deciding whether to flee or freeze when overcome by gas, before the Rotomahana eruption (Star June 17). In 1903 and 1904, the bodies were reinterred, hence in 1905, a new group emerged, apparently in the open, i.e. not indoors. At Waitangi, 23 years later, bodies were recovered in a whare with an intact roof (Herald, 26 July 1909). In 1932, the frames of seven houses and a meeting house at Tarawera stood through seven feet of ash (Waikato Times). Māori homes were robust. If/when further bodies are exposed, they may be found in prone or prone-oblique AJVS 29: 2 (2025) General Issue 36 positions. The residents would lose consciousness in seconds from gas and collapse. The patient does not suffer, unlike death in a volcanic surge with fractures, burns and scattering. Interestingly, the eruption also changed the relative surface and bottom elevations of Lakes Tarawera and Rotomahana, enabling backflow through the porous isthmus between the two lakes, with the Kaiwaka Channel forming one rubble drain. This drainage into the crater occurred for a time during 1887–1893, until the Rotomahana level exceeded that of Tarawera and normal waterflow into Tarawera resumed (Bell, Hodgson and Nairn, Bunn 2022). 3.14. Tokiniho and Waitangi Settlements The 1909 report of 18 Waitangi casualties was checked for confusion with Tokiniho but is likely substantive (NZ Herald 1909), (pers. comm. Rangitihi Pene, July 2025). The Tokiniho toll was first 10, then 13 and by Keam 11 (Keam 1988). The Waitangi survivor Kimihia first reported she and husband Te Hoe were in a large house at Waitangi. The roof fell in, her husband died and she fled to Kariri, being rescued on June 15, en route (Star June 15). Her claim that she took a left turn along the road meant she was heading for Kariri, not Wairoa. All left turns off Spencer Road go to the lake. Later, she changed her account to her aiming for Wairoa and Te Hoe dying en route (Keam 1988, Smith 2011). She reported that at Waitangi ”nobody was left but herself” (Star June 15). In 1909, 18 bodies were found in a whare on a hill near Lake Okareka (NZ Herald 1909). The roof was intact, versus the Tokiniho whare’s collapsed roof in 1887. The body count was 18 versus 10–11. The skulls were counted in 1909 giving a more accurate count for multiple skeletons. Waitangi is closer to Lake Okareka than the estimated position of Tokiniho (800 m vs 2,100 m). It was a large settlement that spread over Spencer Road and up the hill towards Lake Okareka. In 1853, the HMS Pandora crew had embarked from Waitangi. 3.15. Ballistic projectiles The Wairoa eyewitnesses report flaming rocks penetrating homes, igniting fires and injuring residents. This puzzled the author as Te Wairoa is 11–14 km from the Tarawera massif, and the Rotomahana and Okaro vents are 8–10 km away. From Thomas (who also found it remarkable) forward, it has generally been accepted that no molten lava surfaced at Rotomahana, and there were no fresh lava bombs ejected. Finally, Rowe et al confirm that juvenile basalt was ejected from Rotomahana and Okaro as suggested by Nairn in 1979. This helps explain the red-hot rocks falling on Te Wairoa. However, while Te Wairoa is closer to Rotomahana, the distances from Tarawera and Rotomahana are maximal. Taddeucci et al reviewed 28 volcanoes from 17,000 BP– 2014 . Eruptions like Rotomahana-Okaro ejected bombs for < 1 km. Eruptions like Tarawera ejected bombs for 7–10 km, with Vesuvius the only volcano in recorded history to reach 9 km. A. Rex Bunn 37 Fig. 6. A (wet) lava bomb with gypsum and zeolite inclusions, found in the shallows beside Tarata Peninsula in 2015 (Bunn). In Figure 6, is a lava bomb the author recovered from the Tarata Peninsula. It would likely be ejected from a nearby vent. To test the likely origin of red-hot bombs falling at Wairoa, I applied the EJECT v. 1.4 ballistic trajectory model developed by Larry Mastin of the United States Geological Service (USGS). In Figure 7, the EJECT model is applied to the Rotomahana-Okaro vents and Te Wairoa elevations with 0·5 m All-Shaped bombs and a velocity of 400 m/s, as advised by Mastin. The maximum range is barely 6·5 km and less with asymmetric or smaller ejecta. For Rotomahana or Okaro projectiles to reach 8–10 km to Wairoa appears unlikely. A tailwind is a possible reason, but the wind rose and blew from the south before strengthening to the southwest, driving the ashfall away from Wairoa (Thomas 61). AJVS 29: 2 (2025) General Issue 38 Fig. 7. Mastin’s EJECT model for Rotomahana-Okaro vents with All-Shaped ejecta and a maximum velocity of 400 m/s. The maximum range is barely 6·5 km. In Figure 8, the elevations are for Ruawahia vents to Te Wairoa. Here, the plunging bombs reach ~ 8·5 km, still short of the 11–14 km range from Tarawera. The only vent within range of Te Wairoa is the Black Crater at 8·2 km (Nairn 1979). Perhaps there are undiscovered vents nearer Te Wairoa? Ron Keam advised me he thought the Tarawera Rift extended kilometres southwest of Southern Crater at Waimangu. However, this increases the distance to Te Wairoa (pers. comm. Ron Keam, 15 November 2015). A. Rex Bunn 39 Fig. 8. Mastin’s EJECT model for Tarawera Massif vents with All-Shaped ejecta and a maximum velocity of 400 m/s. Maximum range is 8·5 km. 3.16. Logistic analysis Given that some populations survived while others did not, an investigation of the factors affecting survival was made by logistic regression. The first model runs suggested important variables, i.e. Age, Health, Night Vision, Leadership, Flight or Freeze, Path access, Motivation and Distance to vents. Variables with a negative impact were Reaction time, Group size, and Eruption phase. The beta coefficients pointed to Freeze or Flight, Leadership and Night vision being the most important decisions for survival. Reaction-phase timing and Group size were negative, i.e. delaying a decision and increasing group size militated against survival. To improve model performance, the variables were reduced to four: distance from Rotomahana vent, Leadership, Path access, and Decision speed. This model suggested that the four factors explained ~ 40% of the survival of successful groups. However, the result was not statistically significant. To increase significance, the variables were reduced to two, i.e. Distance to the Rotomahana vent and Path Access. This model explained 72% of the survivors’ success and was statistically significant at p = 0·001. The beta coefficients together were significant in predicting the survival choice, but neither variable was significant alone. While useful, the analysis is limited by the small dataset. In AJVS 29: 2 (2025) General Issue 40 this case, it seems intuitively correct that the two most important things in an eruption are to have a path away from it and to obtain as much distance as possible. 3.17. Time and Distance modelling of eruption escape paths The logistic analysis suggested a focus on paths around the three lakes and the window of opportunity for the settlements. For this analysis, the eruption phase start-times are shown in Table 4. The author places greater weight on eyewitnesses whose employment made them timely, such as the tourism guides and Thomas, who surveyed opinions. Table 4. The Eruption Phases From Table 4, observers close to the eruptions cited earlier quake times, for example, Guide Sophia from ~ 11.35 PM. The gap between shaking and the first eruption gave two Golden Hours for escape decisions. Most disregarded the early shaking and lost precious time. As the shaking intensified, those, for example, on Ruawahia exited before the eruption and survived. The first eruption was Wahanga. There was a second Golden Hour for those at Puai and Waingongongo, until the Chasm and Okaro erupted at the other end of the rift. A third Golden Hour occurred until Rotomahana erupted. Based upon survivor and casualty reports, attempts were made by land and water, such as at Moura. After ~ 3.30 AM, there was no escape from the Red Zone. Using time and distance estimates along known paths, we can model the elapsed times from those locations. These can be compared with the windows of opportunity, when flight was feasible. While daytime family walking along familiar paths might average 3·5–4·5 km/h, that winter night with poor visibility, shakes and roaring volcanoes, families might average 1·5–2·0 km/h and less during the shaking. Children would panic, adults freeze, and dogs howl. For this analysis, we use an average speed of 1·0 km/h. An example of the run North from Puai, on foot to Te Wairoa, is in Table 5. Kumete Ridge could be reached, only to be overtaken by the Rotomahana base surges. Eruption Phase Timelines Event Start * Notes Shaking 2330 – 0030 Timings vary by Wahanga 0130 – 0140 observer location. Ruawahia 0200 – 0210 Tarawera 0210 – 0230 Chasm ~ 0230 Okaro ~ 0230 Rotomahana 0320 – 0345 * Times are expressed in 24-hour notation. The major eruptions ended by 0700 – 0800. A. Rex Bunn 41 Table 5. Time and Distance study for those on Puai and Pukura islets. For those taking to canoes, there were also hazards. The quakes would create seiches, which swamped the dugout canoes designed for calm waters. Lake Tarawera is 11 km long, and wind-waves reach hazardous heights along the fetch. On the night, the canoes were overloaded with terrified families who would lie down to cut the risk of capsize. It was usual for passengers to bail in rough weather. A canoe might reach ~ 5 km/h, but if shipping water and overloaded, we use a realised speed of 3 km/h. This is still three times that of foot-travel and would encourage the use of canoes despite the seiches and ballistic projectile risk. At Wairoa, bodies were exhumed along the roads after fleeing folk were felled by projectiles. However, there are no reports of empty canoes drifting in mid-lake after the eruption, nor of canoes reaching Te Wairoa on June 10. After 15 time and distance simulations from key locations, providing families fled during one of the Golden Hours, they could make several kilometres before the Rotomahana surges. By canoe from Oneroa to Punaromia was promising, but no one returned on this route. They elected to stay. No one from Moura survived. The best strategy was to quickly exit Waingongongo and run southeast for Te Whaiti via Ash Pit Rd on Lake Rerewhakaaitu. We know at least one undocumented Māori person survived, and I suggest he or she took this route (Moore 130). 4.0. Discussion Calculating the Māori survivors and death toll is complicated by the ambiguity over locations. Te Ariki was the colloquial Māori name for the Ngutuahi Arm of Lake Tarawera. The maps of Smith and Hochstetter support this. Photographs are often labelled Te Ariki. For example, Figure A5 is labelled “Te Ariki, Head of the Lake”. From this research, there was no single kainga called Te Ariki. The literature records that the major Māori eruption casualties occurred at Te Ariki (27–52 deaths). Te Ariki became the collective term for Kokotaia pa, Piripai and Ngawhiro in Rapatu Bay, Waingongongo and probably the kainga in Te Rata Bay. This research suggests that the cluster at Oneroa was separate. The uncertainty led to understatement of the death toll. Two settlements around the Ngutuahi Arm were renamed Piripai, in Rapatu Bay and Te Rata Bay. In Rapatu Bay, there are different locations. In Te Rata Bay, there is Hochstetter’s location, but no NZAA record. In Oneroa Bay, there is corroborating NZAA garden data and a matching coastline, but poor topography. On balance, and with the inclusion of Ken Phillips' records and Taylor’s sketch, the placename Piripai1 lies in Rapatu Bay. If so, that leaves an undocumented Piripai2 in Te Rata Bay, and Totarariki and Matakana Escape Routes from Puai Islet 10 June 1886 Start Route Waypoint Meters Method Speed Transit Cumulative Cumulative Destination Possible Km to Location km/h Time mins. Distance m. Time mins. Success Locus Puai Shore Tekapo 257 Canoe 3 5 257 5Shore Possible 0·8 Tekapo Path Ngawhana 173 Foot 2 7 430 12 Huts Possible 0·6 Ngawhana Path Cave 175 Foot 2 7 605 19 Shelter Possible 0·4 Ngawhana cave Path Hilltop 135 Foot 2 4 740 23 Eruption view Possible 0·3 Return to Cave Path Cave 135 Foot 2 4 875 27 Cave Possible n/a Discussion Cave 0 10 875 37 Cave Possible n/a Cave-Tekapo Path Tekapo 438 Foot 218 1,313 55 Path to North Possible 0·6 Tekapo-Kaiwaka Path Terrace 1,025 Foot 241 2,338 96 Poroporo crossing Possible 1·2 Kaiwaka-Saddle Path 2,280 Foot 296 4,618 191 Kumete Saddle Unlikely 3·1 AJVS 29: 2 (2025) General Issue 42 undocumented along Oneroa Bay. No survivors are reported from these bays. From this research, Oneroa Bay had the third largest loss of life. It is possible that some Te Ariki photographs are of gardens along Oneroa Bay. The paucity of homes may reflect seasonal occupation. Given Māori proclivity for naming landforms and features, multiple gardens would be named. Watt states the four Oneroa Bay gardens lie near Oneroa. Figures A4 and A5 may show Totarariki and a Piripai. Future archaeological investigation should be made here and at Te Rata Bay. For groups who froze in their homes, it proved successful only in Te Wairoa and Makatiti. At Wairoa, two Māori buildings with steep-pitched roofs, some smaller residences and a henhouse shed the ash. One damaged hotel survived. Barnard notes that Te Wairoa received ~ 80 cm of ash and mud ... with drifts of over 2 Metres ... [and] If 30 cm of ash falls on Rotorua or Kawerau, many buildings will collapse under the load of ash. The 4 kPa standard that most roofs are built to, will be equal to about 250 mm of dry tephra, or 200 mm of saturated tephra... The only possible survival in the Red Zone during the surges would have been in caves. Chief Rangiheuea’s party had that option at Ngawhana. 5.0. Conclusion While most lake settlements are unexcavated, it appears many Māori residents froze at home and died. Flight offered a better chance in the Red Zone and beyond it from vulnerable homes, as Chief Wi Te Kepa Rangipūawhe decided. The two Māori buildings at Wairoa prevented greater fatalities. The investigation into survivors of the 1886 Tarawera eruption is best undertaken now. The lack of online data hindered past investigators. Before the advent of Hochstetter’s survey, there were no reliable spatial data or survey bearings for the Tarawera, Rotomahana or Rotomakariri settlements. Despite generations of geology research, knowledge of the eruption process at Rotomahana was incomplete until 2021 (Rowe et al). The impact of gas emissions was largely overlooked. With these innovations and the NZAA reports of Watt, Moore and Phillips, the Tarawera demographic challenge becomes soluble. The improved understanding of Te Ariki is central. The focus on Te Rata Bay and Oneroa Bay solves demographic questions over Totarariki and Matakana casualties. A similar focus from the author’s Okareka home addresses Tokiniho and Waitangi casualties. The logistic regression was handicapped by the small dataset, but indicated key survival variables. This enabled the Time and Distance modelling that helped explain why (apart from Ruawahia), no survivors from within the Red Zone are reported. At least one did survive, possibly from Waingongongo. It is likely that in pursuing a casualty list, Type II statistical errors occurred— casualties were omitted if a name was in doubt. The Oneroa Bay-TotararikiMatakana and Tokiniho-Waitangi findings and this review increase the casualty list to 172 Māori persons with an upside. Those who froze, if given their time over, would likely choose flight. A. Rex Bunn 43 Works Cited Ansell, Rebecca, and John Taber. Caught in the Crunch: Earthquakes and Volcanoes in New Zealand. Illustrated by Salli Rowe. Auckland: HarperCollins, 1996. Auckland Star. XVII, 137. June 14, 1886. ______ . “The Maoris Lost.” June 16, 1886. _______. “Volcanic Eruptions.” Evening, June 17, 1886. ______ . “The eruption at Tarawera.” September 11, 1886. Auker, M.R. et al. “A statistical analysis of the global historical volcanic fatalities record. J. Appl. Volcanol. 2. 2. (2013). https://doi.org/10.1186/2191-5040-2-2 Barnard, S. C. Potential Physical Effects of Any Future 1886 Type Eruption From Tarawera Volcano On The Bay of Plenty Region. Masters thesis, University of Canterbury, Christchurch, 2003. Bell, J. M. The great Tarawera volcanic rift, New Zealand. Geog. Jnl. 27.4 (1906): 382. Bennett, W. The eruption of Mount Tarawera [Radio broadcast]. New Zealand Broadcasting Corporation. Ngā Taonga Sound & Vision, ref. no. 31288. June 10, 1954. Brabant, H. W. Census of the Māori population, February–March 1886 (Appendix to the Journals of the House of Representatives, H.–3.2.3, No. 17). Wellington: Government Printer, 1886. ______. Reports from Officers in Native Districts. 20 May, 1887. Journals of the House of Representatives, G.–1, Sess. II–1887,1887. Bunn, [A.] R. and N. Drane. “Economic Change as a Factor in Heart Disease.” New Doctor (1978): 53–55. Bunn, A. R. “Ischaemic heart disease mortality and the business cycle in Australia.” American Journal of Public Health, 69.8 (1979a):, 772–81. https://doi.org/10.2105/AJPH.69.8.772 ____. “Unemployment, morbidity, and mortality.” The Lancet, 313 (8122), 923–24. (1979b). https://www.thelancet.com/action/showCitFormats?doi=10.1016:2FS01406736%2879%2991399-0&pii=S0140-6736%2879%2991399-0 ______ . “II. IHD Mortality and the Business Cycle in Australia.” American Journal of Public Health, 70. 4 (1980: 409-11. ______ . Quest for the Pink and White Terraces: Volumes 2 & 3. Bunn, 2020. ______ . “The first evidence-based altimetry for locating the lost Eighth Wonder of the World: the Pink, Black and White Terraces.” Academia Letters, Article 5204 (2022). AJVS 29: 2 (2025) General Issue 50 Fig. A3. Hochstetter’s map of Rapatu Bay showing Kokotaia in the eastern inlet with Piripai in Te Rata Bay. Ngawhiro is unmarked but was east of Kokotaia according to NZAA (From: S. Nolden & S. B. Nolden, Hochstetter Collection Basel: Part 3 – New Zealand maps and sketches, Auckland: Mente Corde Manu, HCB 3.5.6 (detail)). A1.1.2. Te Ariki in Te Rata Bay? Dieffenbach, in 1841, was probably the first trained Western observer in the area. He reports staying in a small bay and bathing in a warm rock pool. He climbed the ridge and took a bearing on a small lake at “south 60°east” (120°) with the two lakes of Lake Rotomakariri. We add 14° declination for a 134° azimuth. From the Wairua Stream, this bearing bisects Lakes Rerewhakaaitu and Rotomakariri/Rangarua. Taking the reciprocal of 314°, his location lay above the Wairua exit and was Otuapane in Te Rata Bay (Dieffenbach 385). In November 1853, the crew of HMS Pandora visited. Lt Jones reported they ... Left Waitangi at 9AM paddled across Lake Taraweraour natives sang the whole way. Peripi [Piripai ARB] just before midnight. There are hot pools there, into which our natives plunged. Next day, walked across the portage to Lake Rotomahana, where the party stayed for two days. This implies that Peripi lay in Te Rata Bay with the portage across the low ridge above the settlement (but it could imply Rapatu Bay). Smith in 1858 also reported a kainga in Te Rata Bay. Smith labelled Hakaipari as the western Ngutuahi Arm and included two sites Hochstetter missed, Pukekiore Pah, above the Kaiwaka Channel exit and a kainga in western Te Rata Bay. The NZAA record identifies Hakaipari as an island pa at Otuapane. There is no sign in bathymetry. It lies under ash eroding down the Wairua Stream. A year later, Hochstetter also reported a kainga in Te Rata Bay named Piripai. His 1859 mapping was assisted by Rev. Spencer over April 31–May 2 as “… heavy rain held him at Te Mu for two more days”. According to Nolden, “…Taking advantage of Spencer’s knowledge A. Rex Bunn 51 of the district, Hochstetter spent the next two days completing his maps ...” Such a sojourn with a local expert must have contributed to Hochstetter’s mapping (Bunn 2020, 31). Rev. Spencer would know of the name change to Piripai, but we do not know whether he saw this map entry. If he did, and it was an error, he would correct it. The author flyfished Te Rata Bay with Cliff Soanes of Rotorua Anglers Association, and he reported a kainga on the ridge. The NZAA log only Pukekiore Hill in Te Rata Bay. Smith mapped Pukekiore Pa in Rapatu Bay. The traditional landowners do not record a Te Rata Bay kainga (Te Arawa Lakes Trust). A kainga in Te Rata Bay was ~ 1 km from the White Terrace (as the crow flies). The Rapatu Bay settlements were ~ 2 km. Guides and tourists would prefer the shorter walk. The Te Rata Bay and Rapatu Bay kainga were probably referred to as Te Ariki. One or both were later termed Piripai. Between 1859 and 1886, there were no large-scale maps of Rapatu and Te Rata Bays. In 1886, Smith showed Te Ariki west of Pukekiore Hill in one map and west of Rapatu Bay in another, with Ngawhiro in Rapatu Bay. That year, Edward Weller (1819 – 1884) showed Te Ariki beside Pukekiore Hill. In 1887, Algernon Thomas (1857–1937) placed it there. Alfred Warbrick showed it at the bottom of the Ngutuahi Arm. Later researchers followed the 1886 maps. The traditional landowners relied on the NZAA for sites in Rapatu and Te Rata Bays (Te Arawa Lakes Trust 2016). These rely on Watt and often do not specify locations. In 2001, Ken Phillips completed the report “Archaeological Survey Tarawera & Waimangu Scenic Reserves.” This includes Te Ariki, with the Māori history concurring that Te Ariki was near Kokotaia, which was abandoned in c. 1855. On balance, the evidence favours Ngawhiro being Te Ariki. Phillips notes John Johnson (c. 1794–1848) in 1847, describing Te Ariki: A few poor huts lying on a narrow strip of land beneath an imposing basalt precipice and on the plateau above formed the settlement of Te Ariki. This latter spot had been selected from having the double advantage of a shelving pebbly shore, where canoes could be beached ... and a wai ariki or hot bath ... This favours eastern Rapatu Bay. Note: Every Te Ariki site has an eruption landform change. The ejecta depth is said to be ~ 9–12 m while the lake level is 6–8 m higher. To an extent, one cancels the other. Given that the period mapping is inconclusive and tourist diaries equivocal, after 1859, we turn to photography. Alfred Burton (1859–1932), William Crawford (1844–1915) and Charles Spencer (c. 1856–1913) exposed Te Ariki views. Watt (and Moore) relied on a Spencer photograph in Figure A4. The author, a diagnostic radiographer, also interpreted this print. The field of view is narrow, implying a long focallength lens, rather than a cropped negative. The ridge looks stacked. The foreground and midground imply tilt. The people are on swampy ground, which was inundated when the lake rose. Across the water are a kainga and a large garden, where the layout matches Figure A6 and other photographs (Te Papa photographs 048755, 0.000969/02, 010293, 042284). AJVS 29: 2 (2025) General Issue 52 Fig. A4. Te Ariki destroyed 10th June 1886 by a volcanic eruption of Mount (Alexander Turnbull Library 1/2-139861-F). Fig. A5. Te Ariki, Head Of Lake. Tarawera. The Te Keu family (Alexander Turnbull Library 1/2-037815-F). In Figures A4 and A5, the ridge above the homes can be compared with a geographic visualisation tool, i.e., Google Earth Pro. In Figure A6, the ridge above Hot Water Beach in Te Rata Bay gives a reasonable match to Figures A4 and A5. A. Rex Bunn 53 Fig. A6. Possible Kainga Location matching Ridge above Te Rata Bay (Google Earth/Bunn). Figures A7 and A8 are a similar view across Rapatu Bay. Figure A8 has a similar background to Figure A4. If it is Te Rata Bay, the camera would be on an islet. If Figure A5 is Rapatu Bay, the camera is below Pukekiore Hill. Across the bay is a point. This is more consistent with Rapatu Bay. If Figures A4 and A5 are Kokotaia in Rapatu Bay, Ngawhiro should show to the east. It does not. Thus, the Rapatu Piripai site appears inconsistent with NZAA records, and this kainga may not be Piripai, at least not the Piripai in Rapatu Bay. Fig. A7. Possible Kainga Location and perspective at western Rapatu Bay (Google Earth/Bunn). AJVS 29: 2 (2025) General Issue 54 Fig. A8. Possible Kainga locations looking south into Rapatu Bay (Google Earth/Bunn). A1.1.3. Te Ariki in Oneroa Bay? Given that Oneroa Bay could represent the Head of the Lake, we consider Oneroa Bay as a site for Te Ariki. In Figure A9, four navigation points match the features in Figures A4 and A5. Fig. A9. The possible Piripai location in Oneroa Bay looking south from the Te Kauae pa shoal, showing features matching Figure A4. (Google Earth/Bunn). A. Rex Bunn 55 These are outlined and include a submerged point, with Oneroa settlement to its south. The shoal is Te Kauae pa. The ridge watercourse exits near the village. The circled peak and skyline match Figure A4. However, the ridge elevation is too high, and Piripai is not here. On a balance of probabilities a settlement in Rapatu Bay, one of several colloquially known as Te Ariki, and later known as Piripai; is the Māori centre with the greatest eruption mortality. For this reason, in the period eruption statistics, mortality at Piripai, Puai Is. and Waingongongo was sometimes aggregated. Within Rapatu Bay, the accepted location during the 20th century came to be on the true right bank of the Kaiwaka at its exit. This was an invention by Keam, possibly based upon a misinterpretation of Figure A4 (Keam 1988). Such a location was improbable for reasons the author provided to the Tūhourangi Tribal Authority in 2017. a) “Soft groundThe alleged location of a Te Ariki village location was on a low-sloping river delta. The ground would be soft and unstable. b) Wet groundThe ground would often be wet or inundated as the lake level rose and fell. c) MosquitosTourist reports cite the Kaiwaka Channel as home to a vicious breed of mosquitos attacking tourists as they boated up and down the channel. These were probably sandflies and/or mosquitoes but these would also have afflicted any Tūhourangi village residents on the Kaiwaka Channel exit. d) Solar anglesA village location under Pukekiore Hill would at some times of the year be late to get morning sun and/or lose it early. The Tūhourangi gardening committee would have argued for their food garden to be beside the village for security and ease of access, but also on a site with maximum solar exposure, to extend the growing season and maximize crop yields. Putting it under Pukekiore Hill makes no sense to a gardener. e) Proximity to shoreThe Tūhourangi villages for which we have photographic evidence i.e. Kokotaia and Piripai, show houses well back from the shoreline. I think this is for at least two reasons i.e. maximum solar exposure and dry feet when the lake rose and fell. f) Poor road accessKokotaia was conveniently positioned on the road from Rerewhakaaitu. This would deliver better communications and trade etc.” (pers. comm. Rex Bunn: Alan Skipwith and Rangitihi Pene, August 2017). https://www.researchgate.net/publication/360560179_A_Note_on_Te_Ariki Acknowledgments The author acknowledges the collaboration with Sascha Nolden, without whom this Rotomahana research would be impossible. Rangitihi Pene has assisted with the research from day one. Thanks to Aimee Foster and Mary O'Keeffe, who assisted with data from the NZAA ArchSite. Special thanks to Bill Fisher, David Lowe, Larry Mastin, Ken Phillips, Mark Rattenbury and the Turnbull Library.