In an age when space probes are routinely despatched to explore other planets, it might be supposed that no new discoveries remain to be made here on Earth. Yet quite the opposite is true. The surfaces of the Moon, Mars, Venus and even the larger moons of Jupiter and Saturn are better mapped than the seafloor of our own planet. A group of scientists from NIWA, the National Institute of Water and Atmospheric Research, is helping fill in the topographical gaps in the New Zealand undersea region. During a voyage of the research vessel Tangaroa in May 2002, the scientists were surprised to discover several large volcanoes, some comparable in size to Mt Ruapehu, between New Zealand and Tonga. Over 14,800 km2 of sea-floor was mapped during the three-week voyage—an area slightly larger than Fiordland National Park. Thirteen new volcanoes of 10 km or more in diameter were found, the largest being 25 km in diameter and over 2100 m high. Numerous smaller volcanoes were also mapped. The voyage was part of a larger project to understand submarine volcanic processes, hazards and seamount biota on the 1000 km-long Kermadec sector of the volcanically active region known as the Pacific Ring of Fire. All the volcanism in the area is associated with the subduction of the Pacific Plate beneath the Australian Plate, which results in the melting of Pacific Plate rocks when they reach a critical depth (about 100 km) and the subsequent rising of that melted rock back to the surface. The some process underlies the volcanism in the Rotorua–Taupo area of the North Island (see New Zealand Geographic, Issue 7). The project dovetails with work on submarine hydrothermal vents by scientists from the Institute of Geological and Nuclear Sciences. A new high-tech mapping system was used to make the survey. A unit mounted below Tangaroa’s hull projected a fan of 135 separate acoustic beams downward through the ocean to map a swath of the seafloor up to 5 km wide. Critically, the system was able to correct for the ship’s motion—its pitching, rolling and heaving—so that the beams could be guided along an exact path on the seafloor. This technology provided a level of detail of seafloor morphology that had previously only been dreamed of, allowing mapping at a vertical resolution of less than 3 m—close to the level of topographic detail on land. Satellite positioning allowed a horizontal resolution of less than 5 m. A major aim of the voyage was to understand the volcanic evolution of Macauley Island, the second largest of the Kermadec Islands. Macauley’s above-water area is less than two per cent of the total area of the largely submerged volcanic complex of which it is a part. The morphology of this complex provides a record of the island’s unusual volcanic history. Magma from the island is, in part, silica-enriched. Such magma was once thought to be associated only with volcanoes built on continental crust, such as those of the central North Island, but Macauley Island is far removed from continental crust, being built on thinner arc crust. Previous studies of Macauley Island have identified silica-enriched tephra or ignimbrite flows which were deposited by a very large explosive eruption thought to have occurred more than 6000 years ago. This eruption was probably similar to the one that created Lake Taupo 1800 years ago. It could well have spewed some 50–100 km3 of material into the atmosphere, and formed a caldera—a giant crater like the present Lake Taupo. Macauley’s caldera is 20 times the size of the emergent island, having a capacity in excess of 17 km3. Swath mapping provides a detailed insight into the Macauley eruption. During or immediately after the main phase of the event, during which the caldera was formed, the volcano collapsed in on itself. This occurs when a volcano has quickly and violently discharged all the material within its magma chamber, leaving the central section of the volcano unsupported. We can tell this happened at Macauley from the pattern of concentric faults on the outer flanks of the caldera. After the collapse, the volcano reverted to basaltic–andesitic volcanism, as shown by rocks recovered from some of the more than 80 small cones that stud the outer rim of the caldera. Lava emitted from these cones flowed both into the caldera and down its outer flanks. Still younger cones are to be found within the caldera. The largest, about the size of Macauley Island itself, rises 250 m from the floor of the caldera, and is sited within 500 m of the island shoreline. Sampling has shown this to be an active hydro-thermal site, and new species of vent organisms, including several species of mollusc, have been recovered from it. These are being studied by experts from the Museum of New Zealand, and are an exciting find for marine biology. The newly discovered vent fauna underscores our limited knowledge of the marine environment even very close to the shoreline, and the importance of marine reserves in the preservation of biodiversity. Both the sea-floor and the overlying water column within 12 nautical miles of all the Kermadec Islands have been set aside as a marine reserve administered by the Department of Conservation. Other caldera volcanoes were discovered south of Macauley Island. One forms a near-circular, 4 km-wide, 500 m-deep hole in the seafloor which looks like a meteorite impact crater. We know that is not what it is from samples of the caldera rock obtained by dredging. These are of normal volcanic composition, and, as in the case of Macauley, concentric faults on the outer flanks of the caldera indicate it collapsed following its formation by an explosive eruption. The caldera wall has been crenulated by small-scale slumping since the eruption, earning the volcano the nickname “Cupcake”. Our mapping results may also provide the location of mysterious Brimstone Island. On September 6, 1825, a Captain Thezar, sailing past Macauley Island, recorded seeing a small, steaming, ring-shaped island 45 km to the west. Later visits by other ships failed to find any trace of the island: it had apparently vanished beneath the waves. For a long time, geologists believed that Brimstone Island must lie on the submarine flanks of Macauley Island—if anywhere. However, we have discovered a completely new volcano 35 km north-west of Macauley Island which rises to within 65 m of the sea surface. Could this be the phantom island? More to the point, is it possible for a volcano to inflate during eruption so that it increases its elevation by 60–70 m and then to deflate? Indeed it is. Onshore lava domes have been observed to grow 100–120 m in height, then to collapse during explosive eruptions. Inflation/ deflation, coupled with erosion by ocean waves, could account for both Thezar’s observation and the island’s current submerged condition. Furthermore, the crest of the new shallow volcano we found is ring-shaped, as Thezar noted of his island. Phases of construction and destruction seem to be the norm for seamounts. Collapse of part of a volcano (called sector collapse) results from the instability of loose volcanic sediment that piles up on the cone: it is readily loosened by the earthquakes which are a feature of tectonically active regions. New research will attempt to establish the frequency of such slope failures, to see whether they cause tsunami. Seabed-mapping technology has come a long way from the humble lead line. Yet today’s scientists are just as awed by discoveries on the ocean floor as their predecessors were. And, like explorers of previous centuries, they must name their discoveries. Eight volcanoes discovered between 1990 and 1994 have, for the most part, been named in honour of deceased New Zealand volcanologists, including Bob Clark, Nick Brothers and Jim Healy. We hope to continue this practice with our new crop of seamounts.