These curves applied to the seasons, if compared with our lava tide applied to the hours of the day, left me with the conviction that the cyclical variations are a fact. They show correspondence between the swelling and shrinking of the globe and the movements of lava, when those movements are free and subject to surveying measurements. For few volcanoes are surveys possible, and our measurements were the first in the world of any continuity. Earthquakes, too, were studied. Dr. Arnold Romberg of the University of Texas--who has become a distinguished inventor in the world of seismology, magnetism, gravity, and oil prospecting--was Professor of Physics at the University of Hawaii about 1918 and for several summers came to the Hawaiian Observatory to assist me in experimental seismology. From 1917 to 1920 I took the records of earthquakes and other seismic movements, as recorded by our Omori instruments, and Romberg remodelled these instruments. With his knowledge of the fundamental mathematics of pendulums, for at Harvard he had experimented with sensitive galvanometers, his facility for making instruments out of nothing but wire, solder, and old clockworks was wonderful and inspiring. I spent many months measuring our smoked-paper seismograms of 1913 through 1918, with the assistance of Mrs. Jaggar, to whom I dictated. I measured types of local earthquakes, of volcanic tremors (some of which definitely accompany lava fountaining), and tilting of the ground, publishing the results in 1920. Tilt upswelling is shown in amount and direction by gradual change of the writing seismograph pens, and this is correlated with the recorded rise and fall of the lava. In the course of three years, with Romberg’s valuable advice, we changed the seismographs to record with little mirrors supported on silken fibers and with beams of light projected on photographic paper. And Romberg invented an ingenious improvement with a vane and a bath of oil, whereby a tilt-free seismograph for earthquakes only would keep the spacing of its lines uniform. Ground tilt crowds the lines. We also experimented with a heavy cylinder which hung as a normal pendulum and which was capable of swinging in any direction, so that it threw a beam of light vertically upward to a chronograph covered with bromide paper. The chronograph was capable of being revolved and stopped, until the mocroseisms and microtremors reached their maximum of amplitude, for any given period of recording. The permanent waviness of ground motion, the tremors with periods of about two-tenths of a second, and the microseisms with periods of about five seconds showed their maxima of back-and-forth movement when the chronograph was revolved to a position where the pendulum swung northeast-southwest. This northeast-southwest tendency was found to be a characteristic of the seismograph cellar for many seismic measurements, including local earthquakes.