June 9, 192 1] 



NATURE 



473 



has been drawn up with the idea of avoiding over- 

 lapping of observation and of including stars the orbit 

 determination of which is hopeful. Many orbits of 

 binaries have recently been computed by Mr. Jack- 

 son, who has also, together with Mr. Furner, pub- 

 lished an investigation showing that the mean mass 

 of binary systems is double that of the sun. Working 

 on this assumption, hypothetical parallaxes have been 

 deduced for several hundreds of stars. The observa- 

 tions with the 28-in. equatorial since 1893 have been 

 collected into a volume, which is nearly ready for 

 publication ; it also contains the orbits found by Mr. 

 Jackson, and notes on the relative motion in cases 

 where orbits cannot yet be determined. 



The programme of parallax determination with the 

 26-in. equatorial is being continued. The plan of 

 taking double exposures on the same plate at six- 

 month intervals has been dropped ; each plate is now 

 developed after exposure. Fiducial plates of each field 

 are prepared bv making rulings with a diamond in 

 the positions of the parallax star and reference stars ; 

 all the plates are compared in succession with the 

 appropriate fiducial plate. Forty-nine parallaxes have 

 thus been deduced in the year, the number of plates 

 measured being 829 ; the probable error of a determina- 

 tion is 0009". 



There are two extensive investigations in progress 

 with the aid of diffraction gratings. The grating em- 

 ploved with the astrographic equatorial gives a first 

 diffracted image 283 magnitudes fainter than the 

 principal image. By successiv^e steps it is possible to 

 compare the magnitudes of all stars within the range 

 of the instrument. The magnitudes of the stars in the 

 Harvard polar sequence are being re-determined. The 

 results obtained so far confirm the Harvard scale for 

 the fainter stars, and the Mount Wilson one for the 

 brighter. 



The grating on the 30-in. reflector is being used to 

 obtain the effective wave-length, and hence to infer 

 the spectral type, of the stars in the Greenwich astro- 

 graphic zone (declination 64° to 90°). An exposure of 

 seven minutes suffices to give satisfactory results for 



stars of magnitude 10-5. Effective wave-lengths have- 

 already been determined for 550 stars within 3° of 

 the Pole, the mean probable error being 10 angstroms. 



The astrographic equatorial will shortly be dis- 

 mounted in preparation for its removal to Christmas 

 Island for next year's eclipse. Sir Howard Grubb- 

 and Sons are making an equatorial mounting for use- 

 there, as the ccelostat method proved unsatisfactory in 

 19 19 for a problem involving such great precision as- 

 the investigation of the Einstein bending of light. 

 Mr. Jones and Mr. Melotte will start for Christmas 

 Island early next»year, and remain six months on the 

 island. The fact of having an astrographic equatorial 

 close to the equator will be utilised for taking series 

 of photographs for the purpose of comparing the 

 magnitude scales of northern and southern zones. 



The Reid and Pons-W'innecke comets have been 

 observed both visually and photographically on several 

 nights. The first photograph of the latter was secured 

 within a few hours of the receipt of Prof. Barnard's 

 telegram announcing his detection of the comet. 



The usual magnetic and meteorological observations 

 have been continued. The mean magnetic declination 

 for 1920 was 14° 87' W. ; it is diminishing by 9I' 

 annually, which will bring it to zero about the close 

 of the century. The chief magnetic disturbance was 

 from March 22 to 25, 1920, being associated with a 

 large group of sun-spots. The mean temperature for 

 the first four months of 192 1 was the highest for that 

 period during the last eighty years, January being 

 7•j^° above the average. The rainfall was 18-77 i"-' 

 being 5-47 in. below the average of seventy-five years, 

 Julv and September alone were above the average. 



The Astronomer Royal refers to the success attained 

 by Mr. Bowyer in the mechanical registration of wire- 

 less signals on a siphon recorder. Signals are re- 

 ceived from the Eiffel Tower, Nauen, Annapolis, 

 Darien, Bordeaux, and Lyons ; some special series 

 were sent from Lyons for the determination of Aus- 

 tralian longitudes. These were recorded both at 

 Greenwich and in Australia. 



A. C. D. C. 



The Chinese Earthquake of December 16, 1920. 



By Dr. C. Davison. 



A 



PRELIMINARY report on the destructive 

 Chinese earthquake of December 16 last has 

 been prepared by Father E. Gherzi, and is published 

 by the Zi-ka-wei Observatory. Though brief, it is of 

 considerable interest, as it is the first scientific account 

 that we have received of this great earthquake. The 

 rep>ort is based on the letters received from corre- 

 spondents of the observatory (nearly all missionaries), 

 on articles in Chinese and other newspapers, and on 

 the seismograms provided by the Wiechert astatic 

 pendulum (mass 1200 kg.) at the observatory. 



The first shock registered there occurred on Novem- 

 ber 16, others on December .', 6. and 10, and possibly 

 three early on December 16. The primarv waves of the 

 great shock arrived on that day at i2h. 9m. 16s., and 

 the secondary waves at i2h. iim. 45s. In less than 

 two minutes later one of the recording levers was 

 dismounted, and after 3I minutes more the other 

 passed off the paper and was put out of action. Such 

 as it is, the seismogcam shows that the epicentre was 

 about 1400 km. from Zi-ka-wei, and that the time at 

 the origin was i2h. 6m. 5s. (G.M.T.). 



The area most strongly shaken lies in the provinces 

 of Kansu and Shensi, in the north-west of the country, 

 in which are situated the origins of the most dis- 



NO. 2693, VOL. 107] 



astrous of Chinese earthquakes. From the somewhat 

 scanty materials at his disposal. Father Gherzi has 

 constructed the probable courses of the isoseismal 

 lines, using the Mercalli scale. The curves of chief 

 interest are those of degrees 10 and <. The former 

 surrounds all the places at which the destruction of 

 buildings was total or nearly so. It includes the towns 

 of Pingliang, Kingchow, Kuyuan, and Tsingning- 

 chow, and covers a district about 180 miles long, 60 

 miles wide, and more than 8000 square miles in area. 

 Its longer axis is directed N.N.W. and S.S.E., and is 

 roughly parallel to the axes of the great crust-folds of 

 this region. Assuming this isoseismal to be drawn 

 correctly, it follows that the position of the epicentre 

 is about 35-8° N., 106-2° E. 



As in all earthquakes of the first magnitude, the 

 duration of the shock was considerable — according to 

 one observer, who measured it. certainly three minutes. 

 Throughout all this time the shock seemed to vary but 

 little in intensity, though becoming slightly stronger 

 near the middle. The effects of the shock were aggra- 

 vated by the structure of the country — the rock in the 

 central area being capped by a thick bed of loess, 

 through which the streams hav'e worn ravines with 

 nearly vertical sides. Roads* are said to be cut up 



