20 



SCIENCE 



[Vol. LVI, No. 1436 



Schumann plates. It consists essentially in 

 dissolving off more or less of the gelatine by 

 means of acid. I have not yet succeeded in ob- 

 taining certain or uniform effects, but in the 

 most favorable cases the sensitivity of the "half 

 tone" plates used in the mass-spectrograph has 

 been increased ten to twenty times without seri- 

 ously altering their other valuable properties. 



The immediate result has been the definite 

 proof of the comples nature of the element tin 

 which had been previously suspected {Phil. 

 Mag., xlii, p, 141, July, 1921). Tin tetra- 

 methide was employed, and a group of eight 

 lines corresponding approximately to atomic 

 weights 116 (c), 117 (f), 118 (b), 119 (e), 

 120 (a), 121 (h), 122 (g), 124 (d) was def- 

 initely proved to be due to tin. This conclu- 

 sion was satisfactorily confirmed by the pres- 

 ence of similar groups corresponding to 

 SnCCH,), Sn(CH3), and Sn(CH3)3. The 

 intensities of the various components indicated 

 by the letters in brackets agree quite well with 

 the accepted chemical atomic weight 118.7, and 

 incidentally preclude the possibility that any 

 of the lines, with the possible exception of the 

 extremely faint one at 121, are due to hydrides. 



The spacing of these eight lines, which are 

 only just resolved, show that their differences 

 are integral to the highest accuracy, but the 

 lines themselves compared with known lines on 

 the plate give atomic weights always tending 

 to be 2 or 3 parts in 1,000 too light for the 

 above whole numbers. That this remarkable 

 divergence can not be explained as experi- 

 mental error is very strongly indicated by the 

 following consideration. The discharge tube 

 had been used previously to investigate some 

 very pure xenon. The line due to Sni-''(CH3) 

 should therefore have appeared exactly halfway 

 between the two strong xenon lines 134, 136. 

 It was actually quite unmistakably nearer the 

 former, so much so that the two were only par- 

 tially resolved. The same irregular grouping 

 repeated itself in another portion of the field 

 in the following spectrum. It seems, therefore, 

 difficult to resist the conclusion that the isotopes 

 of tin have atomic weights which are less than 

 whole numbers by one fifth to one third of a 

 unit of atomic weight, but satisfactory settle- 

 ment of this important point will probably 



have to be deferred till a more accurate mass- 

 spectrograph has been made. 



Incidentally I may add that the presence of 

 the two faint components of xenon 128 and 130 

 previously suspected has now been satisfactorily 

 confirmed. — F. W. Aston in Nature. 



SPECIAL ARTICLES 

 CRATERLETS IN EAST-CENTRAL ARKAN- 

 SAS PROBABLY DUE TO THE NEW 

 MADRID EARTHQUAKE 



The following brief description of six crater- 

 lets occurring on and about the border between 

 the southeast quarter of Sec. 31 and the north- 

 east quarter of Sec. 32, T. 8 N., E. 7 B., is of 

 interest because it apparently extends the 

 sphere of destructive violence of the New 

 Madrid earthquake from that illustrated in 

 Bulletin 494, United States Geological Survey,, 

 to within about 20 miles northwest of Memphis. 

 These are also of interest because few, if any,, 

 larger than these have been described. 



Occurrence: Five of the craterlets occur on 

 the upper surface of one of the low ridges of 

 the Mississippi flood plain. The sixth is a 

 double craterlet, occurring on the slope of the 

 ridge. They apparently bear no relationship 

 to each other except as regards origin and age. 

 With the exception of the double craterlet,, 

 they are saucer-shaped, with diameters ranging 

 from 10 to 40 feet, and depths ranging from 

 2 to 6 feet. One part of the double craterlet is 

 about 15 feet by 10 feet along the diametere 

 of its elliptical outline. The smaller craterlet, 

 is about 10 by 8 feet along similar directions. 

 They are separated by a ridge about three feet 

 high. The bottoms of both craterlets lie about 

 4 to 5 feet below the surrounding surface. 

 These craterlets are all located within a radius, 

 of 800 feet. 



Origin : The ridge on which these craterlets. 

 occur has long been cultivated, and between 

 cultivation and sporadic attempts to fill them 

 up, any evidences of rims that may have sur- 

 rounded the craters have disappeared. How- 

 ever, the material about the craterlets is made: 

 up of the characteristic fine sand and clay that 

 appears in all the dredgings in this vicinity, 

 together with rocks that range in size from. 

 1 inch through to 11x4x5 inches. These roclci. 



