October 19, 1905] 



NA TURE 



do-i 



operator's finger he would be able to feel the pull 

 of the earth's magrietic force upon it. The work done 

 by a few light touches of the finger must be amply 

 sufficient to furnish all the energy required to deflect 

 the needle. But to a modern electrician it certainly 

 seems a remarkably efficient transformation. 



George J. Burch. 

 .\i\ Iiitroducfion to the Study of Colour Phenomena. 



Bv Joseph VV. Lovibond. Pp. 48 ; 10 coloured 



plates. (London: E. and F. N. Spon, Ltd.; New 



York : Spon and Chamberlain, 1905.) Price 5s. 



net. 

 The author states that his object has been to supply 

 the long-felt want of a power of recovering a given 

 colour sensation and of a colour nomenclature by 

 which that sensation may be quantitatively described. 

 To this end " scales of red, yellow and blue were 

 constructed of glass slips, the slips of each scale 

 being all of one colour with a regular variation in 

 intensity from o.oi to 20 units, equal units of the 

 three scales being in colour equivalence with each 

 other. . . . The test of equivalence is that a white 

 light viewed through equal units of the three scales 

 should give no evidence of colour. . . . The fogs on 

 .Salisburv Plain furnished the light actually used." 

 It was found that red, yellow, and blue were the only 

 colours suitable for systematic work, and that any 

 colour could be produced by their combination. The 

 dimensions of the unit are, it is said, necessarily 

 arbitrary, but the scale-divisions are equal, while the 

 unit itself is recoverable. 



The colour to be tested is matched by that of the 

 light transmitted by one of the glasses, or by several 

 superposed, equality of luminosity being secured, 

 when necessary, by the interposition of a neutral- 

 tinted combination between the eye and the coloured 

 object. .A specification of the glasses employed is 

 registered, according to certain rules, as a formula 

 which defines in terms of the author's constants the 

 colour " developed," and supplies data for its future 

 reproduction. 



To those who are accustomed to regard the spec- 

 trum as the natural basis of colour experiment the 

 author's method cannot but appear crude and un- 

 scientific; but, given a sufficient supply of carefully 

 selected glasses, it is probable that much useful work 

 might be done in a rough and ready way by its 

 means, .^n example occurs in the quantitative study 

 of the colour of the human blood in health and in 

 disease, which is illustrated In plate vi. 



The book concludes with an exposition of Mr. 

 Lovibond 's new theory of colour. 

 Index Phytochemicus. By Drs. J. C. Ritsema and 



j. Sack. With introduction by Dr. M. Greshoff. 



Pp. 86. (.Amsterdam : J. H. de Bussy.) 

 l)n. Greshoff explains in the introduction to this 

 volume that it originated in a card index to the 

 literature of plant chemistry compiled for use in the 

 laljoratory of the Colonial Museum at Haarlem, where 

 the work carried on consists principally of the investi- 

 gation of the proximate constituents of plants. 



'I'he index enumerates the names of more than two 

 ihi:usand plant constituents, and gives in each case 

 the percentage composition, formula, melting or boil- 

 ing point, and at least one reference to the literature 

 - — usually Beilstein's "Handbuch," though in a few 

 cases the references are to original papers. The 

 volume also contains a short but useful bibliography 

 of plant chemistry. 



The information given in the tables, so far as can 

 be judged from trials in a few cases, appears to be 

 accurate, and the index should prove useful to 

 chemists engaged in the investigation of plant 

 products. 



NO. 1877, VOL. 72] 



LETTERS TO THE EDITOR. 



[Tlie 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 Nature. 



No notice is taken of anonymous communicaiio)is.\ 

 Eclipse Predictions. 



It is always interesting to compare the results of observ- 

 ation with those predicted by calculation. In the case of 

 the recent total eclipse of the sun this is rendered difficult 

 by the want of agreement in the predictions of the two 

 most used authorities, the Nautical .Almanac and the 

 Connaissance des Temps. The discrepancies in the pre- 

 dicted duration of totality and of the breadth of the band 

 traced on the earth's surface by the total phase are made 

 apparent in the following tabic. It is compiled from the 

 table in the Nautical .Almanac headed " Limits of total 

 phase of the Solar Eclipse," and the corresponding table 

 in the Connaissance des Temps entitled " Limites de 

 I'Eclipse totale et Dur^e de la Phase totale sur la Ligne 

 centrale. " Entries for as nearly as possible the same time 

 in each table have been taken and are placed together : — 



Column A contains the authority. Nautical Almanac 

 (N..A.) or Connaissance des Temps (C.T.). 



Column B contains the time (G.M.T.) for which each 

 prediction is made. 



Column C contains the calculated distance (in nautical 

 miles) and the bearing of the northern limit of totality 

 from the corresponding southern limit. 



Column D contains the durations of totality on the 

 central line as predicted by the one authority and (in 

 brackets) as interpolated from the prediction of the other. 



Column E contains the differences of these pairs of 

 values. 



D 



C.T. 



C.T. 022 ... 113-5 ... N. ' \\"- -..(1984)... 206 ...7-4 

 N.A. 024 ... 101-5 ■•• " 2 W. ... 20O-6 ... (20S) ... 7-6 

 C.T. o 35-2 ... 109-5 ... ,, 2 E. ...(211) ... 219 ... 8-0 

 N.A. o 36 ... 102 ... ., II,, ... 21 1 -S ... (219-5) ... 7-7 

 C.T. 0503 ., 114 ... ,, 6 ,, .. (220-2) ... 228 ... 7-8 

 N.A. 048 ... 104 ... ,, 19 ,, ... 219 I ... (227-4) ■■■ 8'3 

 C.T. I 70 ... 116 5 ... ,, 10 ,, ... (223 8) ... 231 .. 7-2 

 N.A. I 8-0 ... 104 ... ,, 31 ,, ... 223 8 ... (231 -2) ... 7-4 



N.A. 124 ...105-5 J7 „ ■• 22 '-7 ... (226-6) ... 5-9 



C.T. I 24-9 ... 116-5 -•• Ji 1-* >> ■■■ (^20-2) ... 227 ... 6-8 

 C.T. I 43-1 ...115 ... „ 14,, ... (209-2) ... 215 ... 5-S 

 N.A. I 44 ... ic6 ... ,, 44 ,, ... 20S-4 ... (214) ... 5-6 



It will be seen that, for stations in Spain and the adjacent 

 Mediterranean, the duration of totality on the central line 

 was predicted by the French authority to be from seven 

 to eight seconds longer than by the British authority. In 

 the same region, the width of the band of totality is from 

 ten to eleven nautical miles greater by the French than bv 

 the British prediction. The orientation of the line connect- 

 ing the two limits of totality also differs considerably in 

 the two tables. 



It is reported that at Sousse and Gabes, two towns in 

 Tunisia, the eclipse was partial, while a total eclipse had 

 been predicted for them. The prediction for these places 

 would surely rest on French authority : we are therefore 

 entitled to conclude that the mistake has been made bv 

 the French calculators. .\n excessive estimate of the width 

 of the band of totality would almost certainly be accom- 

 panied by an excessive estimate of the duration of totality, 

 and the table shows that both estimates are considerablv 

 greater in the Connaissance des Temps than in the Nautical 

 Almanac. J. Y. Bi-ch.-w.an. 



October 13. 



Absence of Vibration in a Turbine Steamship. 

 Returning homeward to Paris the middle of September 

 from the Tripoli eclipse, and finding passage to America 

 difficult to obtain, I chanced to learn that the triple-screw 

 turbine steamer R.M.S. Virginian was sailing from Liver- 

 pool for Montreal on September 30, so I was very glad 

 to have the opportunity of a voyage in a ship full powered 

 with this novel type of propulsion. .After a week on 

 beard I have no hesitation in saying that for freedom from 



