54 



NATURE 



[September 19, 19 18 



includ ""' ' s main!) pul in thi form 



of tables and curves, and the latter have been repro- 



h dimensions that accurate interpolation 



11 in is possible bj th< use of a rule 



li ■ ini.il parts of a centimetre, ["he prob- 



1 accuracy of data is indicated, or im- 



hi numbei oi significanl figures in the 

 values giv< n 



1 . somewhal surprising to note thai 1 ri"' 1 



has been cast and rolled and afterwards an- 

 led at 500 C. to normalise it does nol have its 

 ultimate tensile strength stated more closely than 

 j5,ooo±sooo lb. per sq. in., when it is remembered 

 that after this treatment there is less variation be- 

 tween different samples than in any other condition. 

 Such coppei has no detectable elastic or proportional 

 limit; i.e. annealed coppei takes a permanent set with 

 the slightest loads. On thi other hand, when it is 

 cold worked, rolled, or drawn, it docs acquire a limit 

 of proportionality, depending on the degree of work. 

 Experiments at the Bureau have shown that modern 

 hard-drawn copper wire is equally affected by drawing 

 throughout the section, and that no hard or exterior 

 skin exists. rhis lias been corroborated by Peirce. 

 Mi, publication is a most useful one. H. C. H. C. 



INTERFEROMETER DETERMINATION OF 

 REFRACTIVE INDICES. 



pROF. CARL BARL'S has recenth developed and 

 *■ extended certain of the methods described by him 

 in iyi6 in connection with the spectroscopic resolu- 

 tion of interferences obtained with interferometers of 

 all classes from the simplest to the most complicated 

 type. 1 Cases of special interest arise in which the 

 interfering spectra are reversed or inverted relatively 

 to each ether. 



Obviously, such methods may have a number of 

 valuable physical applications, and among several 

 examples to which Prof. Barus has given attention is 

 the possibility of the determination of refractivity 

 irrespective of form bv immersion methods. In the 

 method developed for this measurement (chap, iv., 

 part ii.) the interferences produced by white light in 

 a slightlv modified type of Michelson interferometer 

 an viewed with a telescope and prism-grating. Ellip- 

 tical interferences are seen in the spectrum, which 

 may be moved relatively to spectral lines by a micro- 

 metric change of path in one of the beams. A trough 

 containing a liquid of adjustable refractive index, in 

 which the solid under test may be immersed, is placed 

 in this beam, and attempts are made to recognise 

 equalitv in refractive index of solid and liquid by the 

 of fringes on immersion. Naturally, the fringes 

 in the spectrum are distorted owing to unavoidable 

 differeno of dispersion, but it is disappointing that 

 the method should have failed to give a sensitive 

 indication of equalitv. It has, however, long been 

 recognised that interference methods in most cases are 

 inconvenient for direct refractometry ; in reality, the 

 recognition of the point of equalitv is the crux of the 

 matter, for othei more simple and direct methods are 

 available for the measurement itself. In this con- 

 nection an expedient used by the present writer in 

 attacking the probl" Optical Isoc, December, 



10,16) — that of varying refractive index in the liquid 

 bv differential evaporation while homogeneity is 

 secured by mechanical stirring -might possiblv lead 

 to success. The fringes could then be observed con- 

 tinuouslv and the neci ssi tte steps avoided. 



The detection of variations of rei ndex in un- 



Interferometry of Reversed and Non-reverse," - tra." My Car! 



Barus. Farts i. and ii. (Publications of the Carnegie ' •istitution of 



gl ,n. 1916 and 1017.) 



255I, VOL. I02~| 



worked glass is one of the most important problems 

 for modern optics. These variations, which are due 

 to irregularis oi composition, frequently affect the 

 fifth decimal place, but cannot at present be detected 

 until optically worked surfaces have been gi\< n to a 

 specimen. Ii seems possible that the difficult} may be 

 overcome by an intei ferometer-immersion method. It 

 is too often assumed, however, that interferometer 

 methods are of great delicacy in comparison with 

 "definition" tests or the Topler knife-edge test. 

 Remembering Lord Rayleigh's rule, that disturbances 

 should meet in an image with not more than a quarter 

 wave-length difference of phase, it may be realised 

 thai the formation of a well-defined image is a 

 fairly severe test of the homogeneity of the media 

 of the system, having granted sufficient freedom from 

 aberrations due to the form of the surfaces. If, in 

 addition, the direct image is screened so that only the 

 eii, 1 ts of irregularity are perceived, the test ma) 

 apparently be made as sensitive as is desired bj 

 increasing the intensity of the source of light. 



In conclusion, it may be remarked that the method 

 of spectral interferences, although appearing to be 

 exceedingly useful, has not yet been studied so ex- 

 haustively in this or other connections as to enable 

 a final judgment to be passed upon it. L. C. M. 



FERN NOTES FROM PRINCE BONA- 

 PARTE'S HERB AR 1 1 M, 



UNDER the title "Notes Ftendologiques " ' Prince 

 Roland Bonaparte is issuing at irregular inter- 

 vals fascicles of a publication dealing primarily with 

 the fern collections in his private herbarium in Paris. 

 The herbarium already contains about 300,000 speci- 

 mens coming from all parts of the world. These have 

 been derived partly by purchase or exchange from 

 public or private herbaria or from individuals, and 

 partly from Prince Bonaparte's own correspondents or 

 from collectors and travellers with whom he is in 

 touch. Thus many of the collections are represented 

 in other herbaria, and the publication of the nanus 

 of specimens which have hitherto been undetermined 

 will be of service to other workers in the field of 

 pteridology, while a systematic account of new col- 

 lections will add to our knowledge of the ferns and 

 of their geographical distribution. Prince Bonaparte 

 is also pleased to receive on loan collections for deter- 

 mination, and will publish lists of the sp< cies 



The general arrangement is geographical, and each 

 collection is treated separately under the heading of 

 the continent from which it has been derived. The 

 systematic arrangement and nomenclature adopted are 

 those of Christensen's Index. A list of desiderata is 

 printed at the beginning of each fascicli . 



In an introductory note the older practice of relying 

 solely on external characters for the determination of 

 genera and species is adversely criticised. In the 

 future more use must be made of anatomical charac- 

 ters; thus the scales and hairs, which are becoming 

 increasingly important for systematic distinctions, 

 may appear alike when viewed superficially, but on 

 microscopic examination will reveal well-marked 

 characters useful for specific delimitation. These 

 characters, with those of nervation, will be found 

 more trustworthy than those derived from the 

 indusium, a transitory structure. 



These little brochures should prove of considerable 

 value to botanists who are interested in the systematic 

 study of the ferns. We note that the Prince does not 

 follow the rule of giving a brief Latin diagnosis of the 

 new species, though there is often a good description 

 Pari Impr'mi pour l'auteur. 



