What is required from those who undertake to forewarn us of an earthquake is an indication not only of the time at which the disturbance will happen, but also an indication of the area in which it is to occur. Those who dwell in an area where there are certain well-defined periods during which seismic activity is at a maximum—if ten or fourteen days should have passed without a shock—might, in many instances, find that a prophecy that there would be an earthquake within the next few days would prove itself correct. Also, if a severe shock had taken place, a prophecy that there would be a second or third smaller disturbance within a short period would also meet with verification. Certain persons with whom I am intimate appear to have persuaded themselves that they can foretell the coming of an earthquake by the sultry state of the atmosphere or a certain oppressiveness they feel, and an instinctive feeling arises that an earthquake is at hand. It is said that a few minutes before many of the shocks which shook New England between 1827 and 1847 people could foretell the coming disturbance by an alteration in their stomach.[132] No doubt many who dwell in earthquake countries, and have been alarmed by earthquakes, are at times subject to nervous expectancy. The author has had such sensations himself, due, perhaps, to a knowledge that it was the earthquake season, that there had been no disturbance for some weeks, and a consequent increasing state of nervous presentments. In consequence of this, not only has he carefully prepared his instruments for the coming shock, but he has written and telegraphed to friends to do the same. Sometimes these guesses have proved correct. One remarkable instance was a few hours prior to the severe shock of February 22, 1880, when he communicated with his friends in Yokohama and asked them to see that their instruments were in good order. Oftener, however, his prognostications have been incorrect. The point in connection with this subject which he wishes to be remarked is, that the instances where earthquakes occurred shortly after the receipt of his letters are carefully remembered, and often mentioned, but the instances in which earthquakes did not occur appear to be entirely forgotten. He is led to mention these facts because they appear to be an experimental proof of what has taken place in bygone times, and what still takes place, especially amongst savages—namely, that the record of that which is remarkable survives, whilst that which is of every-day occurrence quickly dies. Had the records of all prognostications been preserved, the probability is that we should find that they had, in the majority of cases, been incorrect, whilst it would have been but in very few instances they had been fulfilled.