June 16, 1911] 



SCIENCE 



933 



point in mind, and with a negative result. 

 In response to a query from Professor Doo- 

 little as to the precision of the observations 

 Professor Pickering stated from recollection 

 that the error of a single observation might be 

 three or four seconds, but that the hundreds 

 of observations available brought the probable 

 error weU within the precision required. 



It seems to me there is room for an honest 

 difference of opinion as to the value of the 

 method of least squares in a case like this, 

 where the error of a single observation may be 

 forty-five or more times as great as the quan- 

 tity to be detected. Surely we must draw a 

 line somewhere. 



While I have given this point no extended 

 investigation, I may formulate my own opin- 

 ion, as a basis for discussion, in the form of a 

 mathematical theorem: 



The value of the measure of precision ob- 

 tained by applying the method of least squares 

 varies inversely as the ratio e/q (where e = 

 error of a single observation and q =■ quantity 

 to be measured), in such a manner that when 

 e/q =■ 1 the value is zero, and for e/q > 1 the 

 value is wholly imaginary. 



Paul E. Heyl 



DR. brush's theory op gravitation 

 To THE Editor of Science: The article by 

 Dr. Brush on " A Kinetic Theory of Gravi- 

 tation " in Science of March 10, will become 

 of great interest to physicists when the author 

 does what he partly promises to do in a future 

 paper, viz., explains how a body which is per- 

 fectly transparent to a given radiation can 

 shield another body from that radiation, and 

 why, if the other body is also perfectly trans- 

 parent, it makes any difference whether it is 

 shielded from the radiation or not. It would 

 appear to be immaterial, so far as the effect 

 upon the body is concerned, whether the atoms 

 of a body through which this radiation is 

 streaming in all directions are " buffeted 

 about in every direction by the ether waves 

 in which they are entangled " or whether they 

 remain undisturbed by these waves, so long 

 as they do not absorb any energy from the 

 radiation. 



Dr. Brush says that in the former case, 

 "Each atom or molecule may be regarded as 

 a center of activity due to its kinetic energy 

 of translation, with continual absorption and 

 restitution of the ether's energy normally 

 equal in amount." This seems to the present 

 writer equivalent to saying that a perfectly 

 transparent body may be regarded as one in 

 which the atoms are continually absorbing 

 and radiating equal quantities of the same 

 kind of energy. If anything can be gained 

 by making such an assumption, there seems to 

 be no objection to making it, and I, for one, 

 shall look forward with interest to Dr. 

 Brush's explanation of how it will enable 

 such a transparent body to east a shadow. 



Eernando Sanford 



Stanford Univebsity 



SCIENTIFIC BOOKS 



Soil Fertility and Permanent Agriculture. 

 By Cyril G. Hopkins, University of Illi- 

 nois. Pp. xsiii + 653; 14 illustrations, 3 

 colored maps. Boston, New York and Lon- 

 don, Ginn & Co. Price $2.50. 

 Of this work the author says : " The chief 

 purpose of this volume is to bring together in 

 convenient form the world's most essential 

 facts, gathered from the field and laboratory, 

 and to develop from them some foundation 

 principles of permanent agriculture." The 

 book is a notable contribution to the founda- 

 tions of practical agriculture, treated in an 

 introduction and four parts, I., Science and 

 Soil; II., Systems of Permanent Agriculture; 

 III., Soil Investigations by Culture Experi- 

 ments; IV., Various Fertility Factors. 



The method of treatment adopted is admir- 

 able but not that usually chosen by writers on 

 either soil or agricultural chemistry. The 

 book takes a distinct place in agricultural 

 literature and will be found a mine of infor- 

 mation and valuable reference to the subjects 

 it treats. Professor Hopkins holds per- 

 sistently throughout the volume to the thesis 

 named in the title and does not aim to treat 

 in detail a wide range of topics, but has built 

 his treatment upon a broad, most substantial 



