August 19, 1910] 



SCIENCE 



253 



the seismogram is represented by 15 mm. to 

 the minute, so that individual seconds are 

 easily read. The daily sheets are therefore 

 about a yard long and seven inches wide. 



For the past tvro years the writer has made 

 daily comparisons between the seismograms, 

 barograms and weather maps to find what con- 

 nection, if any, existed between the different 

 phenomena. That on the seismogram we had 

 to deal with earth-movement and not with 

 oscillations of the pendulum becomes obvious 

 to any one examining the record. Undoubt- 

 edly the pendulum may be set in vibration by 

 an impulse, but the former would soon die 

 down on account of the damping, which how- 

 ever is not the case, but instead the pulsations 

 are kept up at times for hours and days. 

 When the period of pulsation is nearly that of 

 the period of the pendulum we find a crescendo 

 and diminuendo showing a spindle shape in 

 the record, and the repetitions of figure take 

 place when the pulsations have gained or lost 

 one on the number of vibrations of the pen- 

 dulum. 



As already intimated an earthquake record, 

 with its various phases of longitudinal, trans- 

 verse and surface waves, can never be con- 

 founded with mieroseisms. 



Let us first compare the microbarograms 

 with the seismograms, taking of the former 

 one that shows rapid fluctuations of pressure, 

 being accompanied by more or less gusty 

 strong winds. On the seismogram will be 

 found simultaneously disturbances, well- 

 marked, but their character is totally different 

 from the microseisms under consideration. 

 Instead of being of the regular form, more or 

 less like saw-teeth, spoken of, they are irregu- 

 lar, a sort of " drunken " record, involving 

 probably undulatory movement beside some 

 pulsations and vibrations. This record dif- 

 ferentiates itself very clearly from that of 

 earthquakes and microseisms. 



Next we compare the weather map with its 

 isobars and areas of high and low barometer. 

 We are soon led to the belief that we are deal- 

 ing in broad outlines at least with two records 

 of the same phenomenon. When in summer 

 the isobars run almost across the continent, 



areas of low and high barometer are iU-defined, 

 and the gradients, or lines at right angles to 

 the isobars, are long, it will be found that 

 the seismograph is almost or wholly quiescent. 

 However, with the advent of autumn the iso- 

 bars begin- to move more closely together, 

 areas of low become confined figures which 

 move regularly across the continent, then the 

 microseisms make their appearance, and seem 

 to some extent to be a counterpart of the 

 atmospheric conditions. We appear now to 

 be on the right track. But a nearer relation- 

 ship is yet to be found. Of one thing we are 

 certain so far, and that is, that if the isobars 

 stretch far over the continent with few and 

 long gradients then there will be no micro- 

 seisms. The converse, however, as experience 

 has shown, can not be so definitely stated, viz., 

 that when isobars are close together, surround- 

 ing a low with steep gradients that then strong 

 microseisms will be recorded. Some micro- 

 seisms will be shown undoubtedly, but not 

 necessarily as a measure of the gradients about 

 a low anywhere. Examination from day to 

 day revealed the important fact that the posi- 

 tion of the low is a very material factor in the 

 production of microseisms. For instance, we 

 have a low approaching over land from the 

 western quadrant, as they always do, showing 

 on the face of it a strong cyclonic movement, 

 yet our seismograph does not seem to be 

 affected by it as long as it is to the west of 

 Ottawa. However, after it has passed Ottawa 

 and descended the St. Lawrence valley or 

 passed along the Atlantic coast to the GuK 

 of St. Lawrence and over Newfoundland to 

 the Atlantic, then the microseisms become 

 very active and we obtain a sheet of serrated 

 lines with amplitudes indicative more or less 

 of the steepness of the gradient. The impor- 

 tant discovery that we have made is that the 

 area of low with steep gradients to be most 

 effective in producing microseisms must be 

 over water, i. e., the ocean. 



So far our facts seem to be well correlated, 

 but there is one essential link still missing, 

 viz., " How does the area of low barometer 

 with steep gradients resting on the water 

 produce the microseisms?" To this question 



