12. Halemaumau, showing lava lake and crags, December 8, 1916
13. Jaggar holding pipe for sounding lava lake, 1917. Cylinder on end of pipe holds Seger cones for measuring lava temperature
By 1918 and 1919 the pit was full and overflowing the Kilauea floor. During the whole of 1919 Halemaumau, as a pit, was obliterated by its dome of fill. In autumn the south flank of Mauna Loa broke out again, into a flood of lava that reached the sea in South Kona. Remembering 1916, we predicted that, even though Halemaumau was full to the brim, the sinking away of Mauna Loa lava would pull down Kilauea lava suddenly, like a siphon. Exactly this happened on November 28, 1919. During the night the crags, the clover-leaf lake, and the bulging dome of the lava fill above Halemaumau’s edge went down as a cylinder to a depth of 400 feet in two or three hours leaving incandescent avalanching walls, a gratifying confirmation of theory.
As in 1916, the Halemaumau lava immediately returned to the bottom of the pit, and lifted itself thirty feet a day for three weeks, so that in December it was a violently boiling ringshaped puddle, surrounding a horseshoe of crags with a quiet inner lagoon and resembling a coral atoll. The Kilauea floor, which is dome-shaped outside of Halemaumau, split open radially to the south, made floods of lava into the Kilauea Crater wall valley, and even escaped out into the Kau Desert. This was extended into a mountain crack, making flank lava flows of Kilauea Mountain, nine miles away to the southwest, something which had not happened since 1823 and 1868. Concentric craters like Kilauea caldera and Halemaumau pit are thus ring-in-ring, or cup-in-cup, structures by means of slag heapings over a deep fracture in the rock crust, the circularity determined by occasional central sinking.
This circularity has sometimes reached perfection. In 1894 and 1909 the liquid pool inside Halemaumau, by steady welling up about a central hole, became perfectly circular within a circumferential rampart of overflow. This is a rare condition dependent on steadiness of upwelling, temperature, and viscosity. It is important because it shows how the perfect circles, and rampart cauldrons, were made on the moon, where there are also angular calderas of subsidence like Kilauea Crater. Evidently gas heating and liquidity changed on the moon, just as it has done in Hawaii. The sources there are over cracks, as in Hawaii. The analogies are so complete in these and many other ways that I completely disbelieve in meteor impact for the moon craters. The moon awaits a complete comparison with active terrestrial basaltic lavas, by a modern volcanologist.
This is only a thumbnail sketch of the astonishing luck which met the photographers and note takers of the Hawaiian Volcano Observatory in its first decade. There were similar decades in the nineteenth century, and there were similar jagged crags rising as islands and shorelines around clover-leaf lakes in 1879 and at other times. There were undoubtedly earlier similar sympathetic movements whereby Kilauea had lowered following the end of Mauna Loa outbreaks. But none of this had ever before been measured from day to day. Our staff from 1912 on occupied the trig stations, every day or night when the weather permitted, in order to measure within one foot the level of the live lava up or down. The lava was like the mercury in a barometer and needed incessant watching. This was done with a telescope, by people who dwelt on the edge of the vertical pipe. After 1913 the measurements clearly showed that sinkings were just as important as risings. They proved that the solid overflow matter and slide-rock slopes around the edges of lava lakes and coves measurably were a paste. This containing bank rose and fell at a different rate from that of the gassy liquid which streamed and fountained inside. The compiled results showed that the source of the liquid streaming was always at the west side of the pit bottom and that the streaming was toward fountaining grottos at the east. The liquid might at any time overflow its banks or sink down leaving inner cliffs, by failure of full supply up the west wall crack.
All of this may sound highly technical; but notes, photographs, seismograms, records of weather, and unceasing press releases and reports to the sponsors, while difficult for literary description, created a new technique. Science, when one is devising a new approach, consists of observation first, of experiment second, and of explanation or theory third. Something of that order has to be followed in the record of a scientist’s life.
The surprising sympathetic lowering of Kilauea following the end of Mauna Loa eruptions was only one of numerous surprises during the first decade of the Observatory. For instance, the temperatures of hot cracks were repeatedly and systematically measured, and nothing sympathetic with lava motion was found. The same may be said about the weather. At the beginning it was supposed that rainfall, air temperature, barometric pressure, and possibly fluctuation of the trade wind, would affect the volcano. However, the only quickly evident effect was the visible vaporing of many cracks on the Kilauea floor which dried up and diminished when the sunshine appeared, becoming dense and increasing in cold or wet weather. This obviously meant that the moisture content of the vaporing cracks, some steam, but mostly moist hot air came from shallow rain water a short distance underground.