^6o 



NATURE 



\_Anonsl ig, 1 8 8c 



of the mountain, by way of Randazzo, Bronte, Aderno, 

 Paterno, and Monte Ste. Anastasia, and so back to 

 Catania. Reflections on the results of the journey are 

 concluded by a perfcrvid peroration, in which the author 

 reminds us that from the top of Etna we may see nearly 

 the whole of that beautiful island which the ancient poets 

 S}-mbolised as " La bionda e leggiadra figlia di Cerere e 

 del sole," and the moderns yet more happily as "la 

 fulgida perla dell' Italico diadema circondata da tre puri 

 zaffiri ; il Tirreno, il Jonio, I'Africano "... 



The book is not illustrated, but it contains a clear and 

 very accurate map of Etna, reduced from that of von 

 Waltershausen, and with the addition of the eruptions 

 subsequent to 1843. Prof. Silvestri's style, while it is 

 accurate and precise from the scientific standpoint, is 

 never dull or lagging. He carries his reader with him, 

 and excites a genuine enthusiasm, which all who know 

 him can well understand. G. F. Rodwell 



OUR BOOK' SHELF 

 Methods and Theories for the Sohition of Problems of 



Geometrical Construction, Applied to 410 Problems. 



By Julius Petersen. 

 Text-book of Elementary Plane Geometry. By the same. 



(London : Sampson Low, 18S0.) 



Some months since we noticed Prof. Petersen's "Theorie 

 der algebraischen Gleichungen," and now we desire to 

 draw attention to two more works by the same writer. 

 The former, in its Danish garb, appeared so long ago as 

 the year 1S66, and having been tried and found to be a 

 successful text-book, the author naturally desired to offer 

 his w'ork to a wider circle of geometers and students. 

 The "i\Iethods"has been rendered also into French; it is 

 "an attempt to teach the student how to attack a problem 

 of construction." Solutions in most cases are merely 

 indicated, the following up the author's remarks being 

 left to the student or teacher. The first chapter treats of 

 " Loci" (method of similitude and inverse figures); the 

 second of "Transformation of the Figures" (parallel 

 translation, replacing, and revolution around an axis) ; 

 the third of "The Theory of Revolution," with an 

 appendix on systems of circles and on the possibility of 

 solving a given problem by the straight edge and pair of 

 compasses. It is a work of considerable merit. The 

 "Text-book" we do not value so highly, though there 

 are points of interest and novelty about it also ; it con- 

 tains besides 228 geometrical exercises. We hail Prof. 

 Petersen as a valuable coadjutor in the work of improving 

 geometrical teaching, and shall be glad if his little books 

 meet with a fair measure of acceptance in this country. 

 We could point out what we consider blemishes, but in 

 the main commend both books. The respective trans- 

 lators (both, we presume, Danish students) have done 

 their part intelligently, and English students will have no 

 difficulty in understanding the language, though they may 

 not be able to master the matter. 



Practical Chemistry. The Principles of Qualitative 

 Analysis. By W. A. Tilden, D.Sc. (Longmans and 

 Co., 1880). 



Of making books on practical chemistry there is no 

 end. If it were necessary that another should be added 

 to the list, the publication of this little book by Dr. 

 Tilden has surely removed the necessity. 



There is no special feature to be noted in this book : it 

 is clearly and accurately written, and proceeds on the 

 well-beaten paths. The adoption of a general table 

 printed on strong paper and protected by cloth backing 

 is to be commended. 



It is, we think, doubtful whether anything is to be 



gained by attempting to teach mere outlines of the 

 methods for analysis of mixtures; a more thorough 

 grounding in qualitative analysis may, as a rule, be 

 given by limiting the student's work for some time to 

 simple salts— which is not such an extremely easy branch 

 of analysis as may at first sight appear ; then proceeding 

 to mixtures of metals with one metal only in each group ; 

 then to mixtures of various metals of the same group ; ancl 

 lastly to complex mixtures. 



The detection of acids— even of a simple acid— is 

 made, as is usual in elementary text-books, to appear a 

 much less difficult undertaking than it really is. 



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. No 

 notice is taken of anonymoits comtmmications. 



[ The Editor urgently requests correspondents to keep their letters as 

 short as possible. The pressure on his space is so great that il 

 is impossible otheiivise to ensure tlie appearance even of com- 

 munications containing interesting: and novel facts.'] 



A Rotatory Polarisation Spectroscope of Great 

 Dispersion 



I HAVE just had an opportunity of trying, on a fine aurora, an 

 instrument for measuring the wave-length of monochromatic 

 light in terms of quartz-rotation of its plane of polarisation. My 

 apparatus is, as yet, very roughly put together, so that I got no 

 measurements of any value, but to-night's experience has shown 

 me that the method, while simple in application, is capable of 

 very great accuracy. 



The construction of the instrument will be easily understood 

 from the auexed rough sketch. The course of the light is with 

 the arrows. N is a Nicol, s an adjustable slit, L a lens at its focal 

 distance from s, Q a plate of quartz cut perpendicularly to the 

 axis, P a double-image prism, and E a small direct-vision spec- 

 troscope, which may be dispensed with when absolutely mono- 

 chromatic light is to be examined. 



When the instrument is properly adjusted by daylight the two 

 images of s formed by P are parts of a straight Hue, so that E 

 gives two spectra side by side. These are crossed by dark bands, 

 which are numerous in proportion to the thickness of Q, and 

 w hich move along the spectra as N is made to rotate. 



In observing a bright-line spectrum the slit is to be made as 

 wide as possible, subject to the condition that no two of the 



N 



7\ ens 



S L 9 P 



differently-coloured images shall overlap. We have thus a pair 

 of juxtaposited rectangles for each of the bright lines, and the 

 angular positions of N, when the members of the several pairs 

 are equally bright, are read off on a divided head. I find by 

 trial that a division to 2° is quite sufficient. 



A first set of readings is taken with a plate Q (permanently 

 fixed in the instrument) 5 or 6 millimetres thick. Then an 

 additional plate of quartz 100 millimetres or more thick is intro- 

 duced between Q and L, and a second set of readings is taken. 

 From the readings with the thin plate we find approximately the 

 positions of the spectral lines, and the more exact determination 

 is obtained from the readings with the thick plate. 



This is the chief feature of the instrument. The actual error 

 of any one reading is not more than 2°, but when a tliick plate 

 is used the whole rotation may be from ten to twenty or even 

 thirty circumferences. By thus increasing the thickness of the 

 quartz plate very little additional loss of light is incurred, while 

 the inevitable error forms a smaller and smaller fraction of the 

 whole quantity to be measured. 



The graduation of the instrument is to be effected by very 

 careful measurements upon a hydrogen Geissler tube, and com- 

 parison with the known wave lengths of the hydrogen hues. 



An observer furnished with this instrument (which is not much 

 larger than a pocket spectroscope) and with a long rod of 



