540 



NATURE 



[August 31, 19 16 



summary of much that has been published in the 

 reports. of the survey. 



Primary, secondary, and tertiary triangulations 

 are included, the permissible triangular error in 

 the first of these being put at 3" and that of 

 tertiary triangulations at 15". The measure- 

 ment of base lines is fully described, and examples 

 are given to show how various sources of error 

 are eliminated. Invar tapes of 50-metres length 

 are used exclusively for base measurement in all 

 grades of work, and a precision of i in 2 million 

 is found to be attainable. The cost of such 

 measurements is given as 2oi. per kilometre on 

 the average, rising to 30L in some cases. The 

 tape is usually supported in the centre and at each 

 end, but in windy weather two additional inter- 

 mediate supports are advantageously employed. 

 The observation of horizontal angles is fully 

 dealt with, and the reduction of the results is 

 explained and illustrated by well-selected examples. 

 A short chapter deals with the subject of map pro- 

 jections, and as this branch of the subject had to 

 be so superficially dealt with, references to works 

 which treat of it more completely might with 

 advantage have been added. 



Two appendices are devoted to the determina- 

 tion of time, longitude, latitude, and azimuth, and 

 to the method of least squares as required by the 

 surveyor. The whole forms a very useful and 

 convenient manual of advanced surveying based 

 on American requirements, but it will be welcomed 

 also by surveyors in British colonies, where much 

 work of this character has still to be done, as it 

 will suggest methods which may suit the cases 

 there occurring. H. G. L. 



The Birds of Britain: Their Distribution and 



Habits. By A. H. Evans Pp. xii + 275. 



(Cambridge : At the University Press, 1916.) 



Price 45. net. 

 Mr. Evans's name is a sufficient guarantee of accu- 

 racy, and this little volume, intended primarily for 

 schools, calls for no adverse criticism. The con- 

 siderable advances in our knowledge of British 

 birds which have been won and "consolidated" 

 during the last twenty years or so have all been 

 taken account of, with due caution as to the present 

 tendency to discover innumerable local forms and 

 to recognise plenty of sub-species. In point of 

 method Mr. Evans adopts a new plan ; he deals 

 with the birds according to their families, giving 

 a separate section to each family, but not to each 

 species. In this way the learner gets a better idea 

 of the British bird-world as a whole, and of the 

 several departments of it, than he could have done 

 from the older books, where the interest was con- 

 centrated on the individual species. No doubt 

 those older books, with their pleasant talks about 

 the ways of a species, will always be both welcome 

 and necessary ; but this one has a value of its 

 own, and is at the present moment the only cheap 

 handbook which is fully up to date. The illus- 

 trations are the least attractive part of it, and 

 much space might have been saved for the letter- 

 press by the omission of some photographs by 

 which nothing seems to be gained. 

 NO. 2444, VOL. 97] 



LETTERS TO THE EDITOR. 



[Tite Editor does not hold himself responsible for 

 i> pinions expressed by his correspundenls. Neither 

 can he undertake to return, ur to correspond with 

 the writers of, rejected manuscripts intended for 

 this or any other part of Nature. No notice is 

 taken of anonymous communications.] 



On Fizeau's Experiment. 



In two papers published in the Proceedings of the 

 Amsterdam Academy (vol. xvii., 445, 1914; vol. xviii., 

 398, 1915) an experimental investigation concerning 

 Fresnel's convection-coefficient for light of various 

 colours was described. The main object of my repeti- 

 tion of Fizeau's experiment, in the improved form 

 introduced by Michelson, was to decide between the 

 expressions for the convection-coefficient given by 

 P'resnel and by Lorentz. As a review of the papers 

 mentioned has appeared in Nature (vol. xcvi., 430, 

 1915),- 1 may be permitted to give here a short account 

 of further progress. It may suffice to recall that my 

 results were largely in favour of the Lorentz ex- 

 pression with the dispersion term. For the wave- 

 length 4500 A.U. the difference between the two ex- 

 pressions under consideration amounted for water to 

 quite 5 per cent. The probable error of the experi- 

 mental result was estimated at somewhat less than 

 I per cent. 



The weak point of my investigation is the deter- 

 mination of the velocity of the running water at the 

 axis of the tubes. This velocity was deduced from 

 the mean velocity combined with the ratio of the mean 

 velocity to the velocity at the axis. The most trust- 

 worthy measurements available at the time gave for 

 this ratio 084, and this number was adopted. A direct 

 measurement of the velocity at the axis would have 

 been preferable, but only lately have I succeeded in de- 

 vising an (optical) method for this purpose. Small gas 

 bubbles introduced into the running water are illu- 

 minated by a very intense, narrow beam along the 

 axis of the tube. A small window in the wall of the 

 tube permits the inspection of the brilliant bubbles in 

 a rotating mirror. From the inclination of the paths 

 of the bubbles, as seen in the mirror, and the constants 

 of the apparatus, the velocity is deduced at once. 

 Direct tests proved the trustworthiness of the optical 

 method. 



Applying this method (Amsterdam Proc, vol.; 

 xviii., 1240, 1916) to my original apparatus, the| 

 window being at the prism end of the arrangement, 

 unexpected results were obtained. The velocil^' 

 actually observed by the optical method not on% 

 diflFered from the formerly accepted value of tlR 

 velocity at the axis by several per cent., but by ne- 

 versing the flow of water its value (at the sani* 

 window) appeared to change by quite 10 per cent 

 Nothing short of a measurement of the velocity at 

 number of points of the tubes and for both directions 

 of the water current became necessary. For this pur 

 pose a Pitot tube, verified by the optical method, wa.' 

 made use of. The results were further confirmed t? 

 the determination of the velocity distribution over di 

 cross-section of the tubes at a few places. Evidend; 

 one cannot speak of the velocity at the axis, as k 

 value changes in a rather complicated manner alonj 

 the tube. A detailed description will be publish© 

 shortly in the Proceedings of the Amsterdam AcadenM 

 The average mean value of the velocity at the aJB 

 comes out 5508 cm. /sec. This is only 5 per ceilf 

 smaller than the value accepted in my principal paptt 

 The conclusions there given remain unchanged, tM 

 thev are now arrived at very directly, all verification 

 of water-meters and the determination of the ra 



