462 



NA TURE 



I March 15, 1906 



itative analysis. The treatment is marked throughout 

 by a considerable degree of originality, and the book 

 appears agreeably free from the domination of an 

 examination syllabus or of the authority of any par- 

 ticular school. It is unusual to find the determination 

 ■of silicon in pig-iron or steel in an elementary book, 

 and so also the use of a Hempel gas apparatus; but 

 there is, after all, no good reason why the practical 

 work of elementary students should not be inter- 

 spersed with exercises of this more technical kind. It 

 'is astonishing what sanctity is still attached to the 

 established order of practical chemistry, and it is not 

 the least interesting feature of this book that it is 

 ■markedly unorthodox. Most teachers will admit that 

 they may profit by carefully inspecting the plans of 

 instruction adopted by their well accredited colleagues, 

 and such a remark may certainly be made of Prof. 

 Hart's little book. A. S. 



La Nature et la Vie. By Henry de Varigny. Pp. 



ii + 356. (Paris: Armand Colin, 1905.) Price 



3.50 francs. 

 In a pleasant and easy fashion the writer of this book 

 carries the reader from the beginnings of life to its 

 termination by death. The origin of life on_ this 

 planet, the vital phenomena of the lower and higher 

 forms of veeetable and animal life, the part played 

 ibv bacteria in the fertilisation of the soil, the evolu- 

 tion of living forms, parasitism, the multiplication 

 of animals and plants, the beginning of the end, 

 the problem of death, and the immortality of the 

 protozoa are a few of the subjects dealt with. The 

 tbook may be recommended as a good popular intro- 

 duction for the educated but non-scientific reader to 

 ^general biological problems. 



LETTERS TO THE EDITOR. 

 [The Editor does not hold himself responsible for opinions 

 expressed by his correspondents. Neither can he undertake 

 to return, or to correspond with the writers of, rejected 

 manuscripts intended for this or any other part of N ature. 

 No notice is taken of anonymo-ns communications.] 



A Plea for Absolute Motion. 



Newton believed in the possibility of absolute motion 

 i(i.e. motion in space not necessarily relative to other 

 material bodies), founding his argument on the fact that 

 •the rotation of a planet might be detected by experiment 

 on the planet itself without reference to outside bodies. 

 Newton's reasoning is unanswerable, but it only takes us 

 part of the way. Though it proves that using the principle 

 of gyrostatic action we can determine direction in space 

 absolutely, it fails to distinguish one parallel line from 

 another. We can only observe relative motion. This state- 

 ment, which no one doubts, is generally taken to be 

 synonymous with the assertion that nothing but relative 

 motion will ever be known. So firmly is this generalisation 

 rooted in the present generation of philosophers that I am 

 afraid the expression of a contrary opinion will only result 

 in placing its author on the " Index " of De Morgan's 

 Budget of Paradoxes. 



It is therefore with considerable hesitation that I venture 

 to raise the question whether we are not most of us in 

 our innermost hearts believers in absolute motion, and 

 whether a good deal of the persistence with which we try 

 in our lectures to prove that no meaning can be assigned 

 to absolute motion does not arise out of the desire to 

 repress our own rebellious doubts. As regards the direct 

 evidence of observation we are all agreed, but if from the 

 outset we limit the results of reasoning to that which may 

 directly be controlled by experiment, we must throw over- 

 board a good many theories which are firmly believed in 

 by men of science. I will try to show that it is almost 

 impossible to exclude the idea of absolute motion from our 

 discussions, and that some of our scientific definitions 

 tacitly admit it. 



NO. 1898, VOL. 73] 



The observed motion of the solar system through the 

 stellar universe has frequently been introduced into the 

 discussion of relative motion, but I do not think that its 

 full importance has been recognised. The thesis I wish 

 to maintain is that the question whether our solar system 

 possesses velocity not only relatively to the stellar universe, 

 but absolutely in space, constitutes a definite problem to 

 which a scientific meaning can be attached. It is 

 immaterial to my purpose whether our present observations 

 are sufficient to allow us to draw any definite conclusions. 

 If the validity of the question itself is admitted, my point 

 is gained. 



In order to free the main issue from the uncertainties 

 arising out of the imperfections of our observations, I will 

 base my argument on an ideal condition of the universe 

 which resembles the real universe sufficiently to be admitted 

 as a possibility. The displacement of a star relative to the 

 solar system may be determined in two ways. While 

 telescopic observations give us the angular motion in a 

 plane at right angles to the line of sight, spectroscopic 

 observations allow us to determine radial velocities. To 

 determine velocities by means of the telescope we require 

 to know the distance of the stars, but the determination 

 of parallax is a question of instrumental perfection and of 

 long-continued observation. We commit, therefore, no error 

 in principle if we imagine the parallaxes of the stars in 

 our ideal universe to be known, so that the combination 

 of telescopic and spectroscopic observations can determine 

 the relative velocity in magnitude and direction. 



It is a matter of history that telescopic observations 

 alone have led to the conclusion that the solar system 

 moves relatively to the stellar system towards a point 

 which, as fixed by Prof. Newcomb's discussion, has a 

 right ascension of 277°-5 and a declination of 35 . Taking 

 this point as apex, Prof. Campbell divided the heavens into 

 eighteen zones, obtained by drawing circles of latitude at a 

 distance of io° with the apex as pole. In every one of 

 the zones which had a smaller apical distance than 90 , the 

 average motion was one of approach to the sun, and in 

 every one of the zones having an apical distance greater 

 than 90 the motion was one of recession from the sun. 

 A complete discussion gave for the line of direction, as 

 obtained by the spectroscopic method, R.A. 277°5, dec. 20°, 

 the right ascension agreeing exactly with the value de- 

 duced by Newcomb, though the declination differs 

 materially. The relative velocity found was about 20 kilo- 

 metres per second. 



We may now idealise this observed universe so as to 

 simplify the argument, and bring out its essential points. 

 Divide the heavens into a number of compartments. Let 

 in each compartment the relative velocities be measured 

 for a large number of stars combining the spectroscopic 

 and telescopic method. Let u be the average velocity of 

 each group relatively to the solar system, so that the 

 velocity of each star in the group can be represented by 

 u + v, both quantities being vectors. For the sake of 

 argument, assume that u is the same for all groups, and 

 that v within each group is distributed according to the 

 law of errors. As regards z>, there is no predominance 

 of any direction (otherwise u would be affected), and its 

 magnitude will be distributed about its mean value in a 

 manner which we will take to be the same for all groups. 

 The question arises : How should we interpret such observ- 

 ations if the facts were as stated? 



It is not sufficient to say that the observations would 

 prove a relative motion —u of the sun with respect to 

 the stellar system, for this would only represent a small 

 part of the facts. The important point brought out by the 

 observations is that the relative motion is observed to be 

 the same for the mean point in each one of a great number 

 of groups of stars. The fact that within each group the 

 distribution follows the law of errors leads to the conclusion 

 that the groups are independent systems, and I put the 

 question thus : Does it require an explanation why all 

 these independent systems should have the same vector « 

 imposed upon them? If you admit the validity of this 

 question, if you begin even to discuss the alternative ex- 

 planation that the vector u reversed really belongs to the 

 solar system, and indicates its velocity, you have practically 

 surrendered to absolute motion. If there were only one 

 star in existence showing relative motion towards the 



