July 20, i8S2] 



NATURE 



2/1 



the minutes can be supplied with perfect certainty by calcula- 

 tion ; but in observations of a transit of Venus the contacts 

 from apparent contact to the " last appearance of any marked 

 "disturbance of the illumination" may, in a moderately bright 

 field, extend over more than a minute of time ; and if any mis- 

 conception of the kind of contact which has been observed 

 should be possible from ambiguity in the description given by 

 the observer, then a serious error may be introduced into the 

 discussion of the results from the adoption of a wrong minute 

 and wrong kind of contact for this observation. In the use of 

 chronometers mistakes of hnlf a minute have occasionally been 

 made by taking the " arrow end " instead of the " longer end " 

 of the seconds hand. 



In all cases, therefore, such precautions should be taken to 

 verify the minutes and half-minutes that errors of entry can be 

 asserted to be impossible. 



19. Approximate latitudes and longitudes of the station, and 

 the authorities from which they are derived, should in all cases 

 be given, together with the local names of the station. 



20. The position of the observer should be permanently 

 marked, and, if possible, referred to three or more surrounding 

 natur.il object":, as mountains, so that the position can be 

 recovered if the mark should be accidentally destroyed. 



21. In cases where the errors of the chronometers or clocks 

 and the geographical position of the observers are independently 

 determined, the observations upon which these determinations 

 rest should be given. 



22. The descriptions of the contacts which correspond to the 

 time records should be written out by each observer, and entered 

 in an indelible form, before any discussion or comparison of the 

 observations with those made by any other observer has been 

 made. On no account is a written figure to be altered. On no 

 account is a new figure to be written upon an old one. Any 

 correction is to be written on another line, and attested by the 

 signature of the observer. 



23. Copies of these observations, authenticated by the signa- 

 ture of the observer, h ith the necessary materials for the deter- 

 mination of dock-errors, longitudes, and latitudes, should be 

 forwarded by the next or following mail to the Committee at the 

 Royal Society, Burlington Hou-e, London. In the case of the 

 Government expeditions, the original documents must be placed 

 in the bands of the official in charge of the Treasury chest at the 

 station, by whom the originals will be retained until the Com- 

 mittee have acknowledged the receipt of the copies and forwarded 

 instructions for the despatch of the originals. 



24. Practice with the artificial models of the transit will be 

 useful to observers as a preparation for the slowne s with which 

 changes in the appearances piesented near the internal contacts 

 take place. But the exact phases presented in the real transit 

 cannot at present be reproduced in the models, and, unless care 

 is taken, model practice may do more harm than good in leading 

 observers to expect a definite succession of phenomena near the 

 internal contacts which they may be unable to recognise in the 

 actual transit. The complicated phenomena presented near the 

 internal contacts are, no doubt, chiefly due to diffractional irra- 

 diation ; but in the case of the models we have the sun and 

 Venus bounded by hard edges. The diffraction phenomena 

 beyond the geometrical boundary of the artificial sun, and the 

 interference phenomena between the limbs of the sun and Venus, 

 are continually changing as the disc, which represents Venus, 

 approaches nearer and nearer the hard edge which represents 

 the geometrical boundary of the sun's disc. These conditions 

 introduce complications into the phenomena seen with the 

 model which have nothing exactly corresponding to them in the 

 real transit ; whilst, on the other hand, the presence of the 

 partial illumination of the atmosphere of Venus introduces 

 difficulties in the observation of the real transit which have 

 nothing exactly corresponding to them in the models in ordinary 

 tie. 



ATOMIC WEIGHTS 1 



SEVEN years after the publication of the first volume 

 of Daiton's " New System of Chemical Philosophy,'' 

 and therefore at a time when the data from which atomic 

 weights could be deduced were few and inaccurate, Prout 



1 " The Constants of Nature. Part V. A Recalculation of the Atomic 

 Weights." By Frank Wigglesworth Clarke. SB.. Pr fess>r of Chemistry 

 and Physics in the University of Cincinnati. (WashinstDn : Smtths:n;an 

 Institution, 1882.) 



promulgated the hypothesis that the atomic weights of 

 all the elements are multiples of that of hydrogen. 



This hypothesis was soon shown to be without founda- 

 tion in fact, but in the modified form given to it by 

 Dumas — viz. the atomic weights of all the elements are 

 whole, half, or quarter multiples of that of hydrogen — it 

 found very considerable acceptance among chemists, 

 although it was strongly opposed by many. 



In i860 Stas published the results of very carefully- 

 made determinations of the atomic weights of nitrogen, 

 chlorine, sulphur, potassium, sodium, lead, and silver. 

 Stas concluded from these results that Prout's hypothesis 

 is purely imaginary ; that each elementary substance is a 

 distinct entity, and exhibits no simple mass relations with 

 other elements. 



Marignac criticised the numbers obtained by Stas, 

 objecting that unless an atomic weight is determined by 

 wholly different series of experiments it cannot be 

 accepted as final, and making the somewhat astonishing 

 statement that possibly the composition of a given com- 

 pound is not altogether invariable. The reply of Stas 

 appeared in the form of his famous "Nouvellesrecherches 

 sur les lois des proportions chimiques, sur les poids 

 atomiques et leur rapports mutuels," wherein the fixity of 

 composition of many compounds was firmly established, 

 and numbers were deduced, from widely different and 

 most carefully conducted series of experiments, for the 

 atomic weights of silver, iodine, bromine, chlorine, sulphur, 

 nitrogen, lithium, potassium, sodium, and lead, which 

 numbers appeared finally to negative the hypothesis of 

 Prout, even in the form given to it by Dumas. 



The experimental work of Stas has been accepted as 

 unimpeachable by every chemist. The " Nouvelles re- 

 cherches " is a classical work. But in 187S Dumas 

 showed that pure silver, prepared by the method adopted 

 by Stas. gave up weighable quantities of oxygen when 

 heated in vacuo. The numbers given by Stas as the 

 atomic weights of the elements enumerated above may 

 therefore not represent the true atomic weights of the = e 

 elements. The importance of the discovery made by 

 Dumas is emphasised when we know that the atomic 

 weight of silver is a fundamental number, on which most 

 of the other atomic weights determined by Stas depend. 



In 1S72, Ciookes communicated to the Royal Society 

 the results of an extremely careful determination of the 

 atomic weight of thallium ; the mean number obtained, 

 203 - 642, was regarded by Crookes as strongly against 

 Prout's hypothesis. 



Recent work, physical as well as chemical, has again 

 caused attention to be turned to the hypothesis which 

 would regard the elements as forms of one kind of matter. 

 The necessity for a revision of many atomic weights 

 has impressed itself on chemists; and several very careful 

 revisions, notably that of the atomic weight of antimony 

 by Cooke, and of aluminium by Mallet, have recently 

 been made. But in addition to these new- data there 

 exist many determinations, which, if properly collected 

 and digested, would be of much importance. Prof. 

 Clarke has done this most admirable service to s 



"Atomic Weight Determinations; a Digest of Inves- 

 tigations published since 1S14." by Prof. G. T. Becker, 

 has already appeared as Part iv. of the Smithsonian 

 Constants of Nature ; Prof. Clarke's Recalculation com- 

 pletes the Digest ; together these form a contribution to 

 chemical science of the first importance. 



The ratio between the atomic weights of oxygen and 

 hydrogen is that first discussed. Each series of experi- 

 ments is considered separately j the mean value is found 

 and the probable error of this mean is assigned by the 

 method or least squares. Those elements, the atomic 

 weights of which have been most carefully determined, 

 viz. silver, chlcrine, bromine, iodine, potassium, sodium, 

 and sulphur, are next considered. 



The discussion of atomic weights involves many chemi- 



