SPANISH LITERATURE

vhe dominant feature of spanish contemporary litcrature in 1910~26 may be found in a more conscious realisation of the creative element of the race by its critical element and, therefore, in a tendency towards the fusion of both these elements into a complete whole. with the death of galdos, the sceptre of spanish literature passes to miguel de unamuno, who represents the modern version of the spanish mystic writer. his main concern is the relation of man to creation. it is the subject of his masterpiece del sentimiento tregico de la vida (1913), a book of passionate meditation (of which an excellent translation exists in english, the tragic sense of life, 1921), and as an attitude of mind it dominates his criticism; zin torno al casticismo (1902), ensayes (1916), la agonia del cristianismo (1926); his novels: niebla (1914), abel sdnches (1917), tres novelas ejyemplares y un prelogo (1920) and his play: fedra. in these works uhamuno appears as the apostle of an ideal of life more closely connected with spiritual fasternism than with the intellectual and social tenets of the west. he thus fulfils in spain much the same function as dostoievsky did in russia, for spain like russia is a transition between east and west. he aims at intensity rather than extension. his style is not unlike the stvje of carlyle in that it is written with the whole being of the man, body and soul. if, in order to complete the parallel with russia, a spanish western type of mind had to be opposed to unamuno, as tur- geniev could be opposed to dostoievsky, a younger man, jose ortega y gasset, might be selected for this purpose. ortega y gasset is a refined humanist, strongly influenced by german contemporary neo-kantian schools of thought. his main work is in the field of criticism and psychology—medituciones del quijote (1911), el espectador (1917). in more recent books he has endeavoured to draw philosophical and psychological con- clusions from current events; for instance, in the field of politics with his espatia invertebrada (1922) and in that of philosophy and science with £7 tema de nucstro tiempo (1923), a masterly commentary on the new vistas opened out to thought by the discoveries of einstein. in this same school may be included jose martinez ruiz (b.1876), better known under his literary name of azorin, whose art has all the finish and exquisiteness, all the smallness also, of miniature painting. he has had the rarer merit of applying these gifts to the interpretation of national scenes and places, as in his castilla (1920); los puchlos (1905); la ruta de don qutjote (1905), and has thus contributed tn no small measure to the movement for national self-knowledge which is noticeable in contemporary spain. fiction.—in fiction, though belonging to an older generation, vicente blasco ibafiez must be mentioned. his war novel spectroscopy los cuatro jinetes del apocalipsis (1916, eng. trans. 1918) has made him famous with the english-speaking public. older novels (and better ones) have been translated, suchas la barraca (1809); the cabin (1919); sangre y arena (1908) and the afatador (1918). blasco ibafiez represents an art which is spanish only in its subject but not in its spirit, manner or style. of a vounger gencration, pio baroja (b. 1872) is perhaps the most widely read. a basque, with all the acuity of mind of his race and not a little of its rustic independence and antago- nism to civilisation, baroja writes abundantly and carelessly with more spirit than art. his best work is perhaps [dzlios vascos, in which he has rendered the quaint charm of his own country. many of his works have been translated into english by ameri- can publishers. to this generation belongs also ricardo leen, who writes in a more consciously traditional vein and pays con- siderable attention to matters of style. his main work is casta de hidalgos (1908). concha espina, a brilliant woman novelist, has distinguished herself by novels of psychological insight and easy style, such as el metal de los muertos (1920). ramen perez de ayala (b. 1881) is perhaps the best novelist of the younger generation. a critic of great talent and a poet, he has written several novels, the best of which are novelus poemdticas de la vida espajola (1916) (translated into english under the title of prometheus) and belarmino y a polonio (1921). he is a typical exponent of the fusion of the traditional spanish spirit with the conscious knowledge of its resources referred to above, as characteristic of contemporary literature in spain. the same may be said of gabriel mire, whose delicate sensibility and deep knowledge of the language make him one of the most richly endowed authorsof contemporary spain. among his works are fl humo dormido (1920); el libro de sigiiensa (1921); figuras de la pasien del sesor (1916) (this last translated into english 1924). of recent years several authors have cultivated the utopian and satirical variety of novel-writing. ‘to this kind belong el archipielago maravilloso (1923) by luis araquistain, and la givafa sagrada by s. de madariaga (originally written in english and published 1925). among newcomers in novel- writing claudio de la torre and felix urabayen, e/ barrie maldito (1925) must be singled out. drama.—jacinto benavente still dominated the spanish theatre during this period. his most famous play, los intereses creados (1907), is not representative, for it illustrates but one phase of the talent of this many-sided author. a more power- ful tragedy, la noche del sebado (1903), is of the same period. in more recent times he has given an intense drama of life in la malquerida (1913). there is, however, a type of play in which benavente must yicld the prize to the brothers alvarez quintero (serafin, b. 1871; joaquin, b. 1873). as authors of comcdias de costumbres these two writers, who always work together, are unsurpassed. the list of their comedies is long, and includes las de cain and puebla de las mujeres (1912). other play- wrights of note are linares rivas (b. 1866), remarkable for his skill in the handling of dialogue; martinez sierra, a name which stands for the collaboration of don gregorio martinez sierra (b. 1881) and his wife, resulting in a happy blend of dramatic skill and delicate psychology; and pinillos (parmeno) (1875- 1923), a vigorous painter of social conflicts. the poetical theatre still lives in spain, maintained by valle-inclin (b. 1870), marquina cata and ardavin. in recent years a writer of un- usual power, jacinto grau, has conquered the foreign stage, particularly with his ei conde alarcos (1917), el hijo predigo {1918). but drama and comedy are only one, and that not the most important, aspect of the spanish theatre. still more typical of the nation is what is modestly known in spain as genero chico (small genre), a type of theatrical production con- sisting in plays generally short and accompanied by music and ranging from variety pieces akin to operettas to little master- pieces of musical drama. its best-known exponents are the brothers quintero and carlos arniches (b. alicante, 1866). poctry.—the two main influences acting on spanish poctry towards the close of the roth century, e.c., national tradition and the example of foreign poetry, particularly that of the symbolist 621 school of france, which reaches spain through south-american poets, such as ruben dario, are still at work. though the first is the more vigorous and conscious, the second widens so as to include all influences: d’annunzio, maeterlinck, tagore. more typically national are unamuno, rosario de sonetos liricos (1911), bl cristo de veldsques (1920); antonio machado (b. 1875). whose pessimistic serenity is in keeping with the landscape of central spain (soledades, campos de castilla) and salvador de madariaga (b. 1886), whose romances de ciego (1922) re-state in a new form the old spanish theme of jorge manrique. other poets produce under more complex influences. thus manuel machado (b. 1874), whose main inspiration is popular and southern has, nevertheless, written excellent verse in which the influence of french elegant sensibility is discernible; juan ramen jimenez (b. 1881), remarkable rather for exquisite sensitiveness than for power (arias tristes and elegias, 1908- 10), is led by his melancholy moods towards fluid rhythms which, though more settled, are reminiscent of maeterlinck and, through him, of rossetti. ramen del valle-inclen (1870), per- haps the most skilful musician among modern spanish poets, has given in la marquesa rosalinda (1913) an admirable example of the adaptability of the spanish language to the most exquisite and complicated rhythms. ramen perez de ayala in his three volumes, la paz del sendere (1916); el sendero in- numerable (1916); and el sendero andante (1921), effects a happy wedding of thought with harmonious poetry in a work not wholly uninfluenced by francis jammes, d’annunzio and walt whitman. among the rising generation may be mentioned pedro salinas, jorge guillen, garcia loria and rafael alberti. amiscellaneous.—in the field of erudition and literary history the task of menendez y pelayo is continued by don ramen menendez pidal, whose works on the poem “mio cid” (1911, 1913, etc.) and on the spanish chronicles have thrown much light on the origin of spanish epic poetry. francisco rodri- gucz marin (b. 1855), editor of don quijetc, a specialist of spanish folk lore, has succeeded the master as head of the national library. father asin has won world-wide fame by his illuminating work on the arabic origins of dante. of the younger gencration federico de onis (b. 1885) has edited fray luis de leen (1914) and americo castro (b. 1885) has worked on lope de vega (19109). journalism, always a great art in spain, where the news- paper is infinitely more read than the book, is cultivated by all writers and everyone of the names mentioned above might be quoted here again. mention must be madc, however, of two eminent contemporary writers whose work is almost exclusively journalistic: ramiro de maeztu (b. 1874), a versatile mind whose educating influence on the spanish reading public has been incalculable; and luis araquistain, a powerful dialectician and a master of the polemic style. braitograpiy.—cesar barja, liferatura espaiiola, libros y autores modernos (1924); fe. l. gomez de baquero, el renacimiento dela novela fespatiola en el siglo xtx. (1924); miguel de unamuno, en torno al casticismo (1902); w. starkie, jacinto benavente (1924); s. de madariaga, the genius of spain (1923). (s. de m.) spectroscopy (sce 25.619).—<as developed since 1910 the science of spectroscopy has for one of its chief purposes the analy- sis of spectra, and the deduction therefrom of the structure of the atoms and molecules which generate the spectra. the prog- ress which has been made in this connection has depended upon improved determinations of the wave-lengths of spectral lines, the further investigation of the varying spectrum of the same substance when excited to luminosity in different ways, the more complete analysis of certain spectra into regular series systems and, finally, on theoretical investigations. in another direction, important advances have been mace in the interpre- tation of the spectra of the various classes of celestial boclies, which may be regarded physically as experiments on large masses of matter at various high temperatures. standards of wave-length.—fxtensive interferometer determina- tions of wave-leneths in the arc spectrum of iron, based upor 6438-4696 “ international ” angstrems for the red cadmium line, 622 have been made by k. burns! and others, which provide the necessary standards for the general determination of wav e-lengths by interpolation. it has been found, however, that the wave-lengths of many lines differ considerably in different parts of the arc, so that special precautions are necessary in order to obtain comparison spectra in agreement with the tabulated payne _ probably the most accurate set of standards are those given by c, e. st. john and hl. d. babcock,? who used a small central zone of a “ pfand ” iron arc opcrated between 110 anc 250 volts with five amperes or less, at a leneth of 12 millimetres. this list contains 1,026 lines, from ’3370 to 46750, and for most af them the wave-lengths are believed to be accurate to o-ool‘angstren. flame, arc and spark spectra.—the range of spectroscopic research has been almost indefinitely extended by the discovery that the same substance vields different a eee when stimulated in different ways. such differences are of little i importance from the point of view of chemical analysis, but they have become of great significance to the physicist, and have also greatly aided in the interpretation of the spectra of celestial bodies. the three typical methods of seadicie luminosity for the ob- servation of the spectra of metallic elements or their salts are the flame, the electric are and clectric spark, as a gener al rule, aes three scurees exhibit important di'ferences. in the flame the line are comparatively few in number; in the are the flame lines remain promincnt, but many more rnes, inctuding some which are as strong as the name lines, make their appearance. in whe spark, there is a tenilency for many of the ty pical arc lines to disappear, whilst other jines may be miuten intens sified, an } entirely now lines may also be present, phe import: ant class ef jines which are la- tensified, or w hich only appear under the violent action of the con- densed spars, were deste rated cnhaneed fines by sit norman lockver, and thisname hasbeen venerally adopted. the clifferent classes of lines are thus commenly known as flame, arc end spark (or enhanced) lines, eens to their relative preminence in these three sources. this ¢ iil classification, however, 13 some respects imperfect, a and more definite designations are now based upen the theaeretical considera- tions to which reference will be raade later. ua detailed temperature. classification of the lines of many clements has been lased upon experizunts with the electric furnece by kk ing. ? puuular \ rariations have also heen observed in the spectra of gases when submitted to the action of discharges cf varying intensity, ancl the different classes of lines are sometimes distinguished by analogy as are and spark lines, although wilh bi exceptions the are is mot actually employed. the actual spectrim ¢iven 1 by a gas donends upon its pressure as well as upon the ae asity of th ct hscharoe by which it is made luminous. generally speaking, the gre ater ve pressure of the gas the pleat will be the streneth of the dischary required to pr duce the “ spark " lines. one important airn of spectroscenic research has been to search for an explanation of these phenomena, for it cannot be doubted that the causes of the variations in the spectra are # dannately connected with atomic structure. in this connection it will be instructive to refer first to the spectra sa known compounds. there are many compounds which can be excited to luminosity without total decomposition,’ and it has been found that each compound gives a characteristic snectrum by wiich it can be identified as such. these spectra inva ariably c consist of bands, and different sets of bands characterise, for example, the cxicdes, chlorides and fluorides of the alkaline carth elements. it is sufficiently obvious that if a compound be stimulated so strongly that it becomes dissociated, the spectrum wiil c from one consisting of bands representative of the compound to one containing the lines of the constituent elements. it ts nut only compounds, however, that show changes of this character. ixxperiments on nitrozen, for instance, show a range of spectra from one consisting wholly of bands to one in which lines occur alone. even hydrogen has twe spectra: (.:) the highly complex, so-called secondary spectrum, which doubtless represents a banded spectrum of rather coarse structure® and () the familiar line spectrum, constituting the balmer series. similar results have been obtained for many other elements, and from analogy * lick obs. bull., no. 247 (1913); zett. f. wiss. phot., vol. 12, 209. 2 astrophys. jour., vol. 53, 260 (1921). trans. int. astron. u ion, vol. 1 and 2 (1922 and 1925). 3 several papers in the astropnys. jour. ‘stimulation by ‘ active nitrogen,” according to the methodls of r. j. strutt (now lord ray leigh) is particularly effective for the spectra of many compouncds.—proc. roy. soc., vol. 86, 105 (1912). § an excellent photographic map of this spectrum has been given by t. r. merton, proc. roy. soc., a, vol, 96, 382 (1920). 2 a 4 sis. mange a spectroscopy with the spectra of compounds the natural conclusion is that the band spectra of the elements arise from molecules, while the line spectra are produced by the atoms which are set free when the molecules are dissociated. if this be a true view, the change in the structure of the atom, or in its mode vibration, which accompanics the successive modifica- tions of the line spectrum becomes a question of paramount interest. lockyer® did not hesitate to believe that while the are lines of an clement were to be attributed to ordinary atoms, the enhanced lines could only be preduced by the splitting-up of the atoms themselv es, ant he called these simpler forms of mattcr the proto-clements. proto-calcium, for instance, denoted calctum which had been broken up into sub-atoms by the application of a sufficient stimulus. a somewhat similar, but more probable, explanation has been based upon an application of the quantum theory by bohr to ruther- ford’s nucleus theory of the atom. this theory is founded largciy on the analysis of spectra into regular series. (see atom.) range of observations —for the complete determination of the ee of spectra it is necessary to extend the observations far hcvond the limits of the visible sp« ‘ctrum. conspicuous success in the direct photography of the pe infra-red spectrum has been achieved by meegers and others, | yy the use of ordinary plates stained with dicvaain.? by this me ‘thod excellent phote ovraphs cf the are spectra of a large number of e! cinents, extending to a10,000, have been ob- tained with a concave grating, and the positions of the lines have been measured with a high degree of accuracy. for the present the extreme tufra-red can only he ins restigated by thermal effects, in- volvi: i the use of the thermonile,' bolometer er radio-mic rometer, as in the researches of paschen, lehmann andl randall. spectroscopic ghee ‘vations in the direction cf the citeantolet ee ail t! he limit about ban 50 set by the abso: ‘ption of quartz, and bevond about atzoo set by the absorption ef air, which were first made by victor sc! mma, ne e he en orcatly exte nde by the use of concave gratings in vacuum spectrographs “by lyman and others. ae milikan’ and his colleagues have observed lines as far as a036 a. several improvenients ti lechnique were necessary to this ss. it wasachteved, in the first place, ‘by using gratings spcctaily hip yted for the purpose; ¢ secondly, by working in an cesential! iy per- fect vacuu m, turoush the use of power ul pumps; and finally, since no ordinary spark could pass 11a vacunm, by the use of a specially strong sparking apparatus which was capable of forcing a discharge across a very small space between the electrodes. spectrosc copic data thus cover a. very wide range, and offer many intere sting problems to the investigator, their solution depends on his ability to make a true analysis of spectra, and to dd: educe there- from the corresponding atomic or molecular conilitions, analysis of specira—great progress has been made im the investigation of the regularities in spectra. in these inquiries the position of a spe ectral line is most conveniently represented by its “ wave-number ” (y); f.e., the number of waves per centi- metre 72 tacuo. in practice it is obtained by dividing 10° as he number of angstrem units in a centimetre) by tite wave-length as corrected toa vacuum. hesimplest spectrums that of ee ‘et, which may be accurately represented by the sim ple fornia i t u= r (— ro ll fits where #t,72 are integers and r is the “ rydberg constant,” hav- tus th 1e value 109,678-3.9 tf n= ancl a takes successive inte; ral values 2, 3, ... the formula represents a series of lines in the extreme ultra-violet which has been called the “sf lyman series.” with #j=2, #=3, 4... ., there results the “ balmer seve extending through the visible spectrum to the near ultra- -viclet. other s scrics which appear in the infra-red, including the “ pas- chen series,” are similarly represented by putting = 2, 4. and so-one. kr, w. wood! has extended the balmer scrics to #1 = 22 by experiments with long vacuum tubes, while 34 members ol this series have been recorded in the spectrum of the sun's chro- mosphere during total eclipses. in each scries, it should be noted, the lines gecome closer together and diminish in intensity as they approach the definite limit indicated by the first term of the formula. theoretically, the number of lines in a series is infinite, but tt often happens that only a few of the earlter mem- bers are observed, and no serics has been traced to its limit. q 3 6 sir j. n. lockyer, inorganic evolution (1900). 7 scientific papers, bureau of stundards, washington, no. 312 (1918) and subsequent papers. = phys. rev., vol. 23 (1924). iw e. curtis, proc. roy. soc., a, 96, 147 (1920). 10 proc. roy. soc., a, 97, 455 (1920). spectroscopy there is only one other known spectrum which has the same simplicity as that of hydrogen—namely, the enhanced spectrum of helium. this includes the series first found by pickering in the star ¢ puppis, and the line x 4686 and others calculated by rydberg, by whom both series were attributed to hydro- gen. these lines were produced in the laboratory by a. fowler,! and additional lines of the pickering series, first indicated by bohr’s theory, were afterwards observed by e. g. evans? and by paschen.? it was, in fact, the theoretical work of bohr which first suggested that the lines in question originated in helium and not in hydrogen. the enhanced series of helium can be represented by a formula similar to that for hydrogen, with the diiference that the series con- stant has rather more than four times the value for hydrogen, namely 4109723. when m=3, and 1 =4, 5... we obtain the series hav- ing a4686 for its first number. the complete pickering series ts given by putting ni =4, m=5,6 ... ,andaseries in the extreme ultra- violet observed by lyman is siven when m.=2. it should be noted that alternate lines of the pickering seri2s are nearly coincident with the balmer series of hydrogen. the spectra which are next in order of simplicity are the arc spectra of the alkali metals, lithium, sodium, potassium, rubid- ium and caesium. in each of these there are severe! serics superposed, and all the lines are doublets. in the alkaline earth metals, beryllium, magnesium, calcium, strontium and barium, there are two systems of series, one consisting of singlets anid the other of tripicts. as in the spectrum of hydrogen, the lines of each serics converge to 2 definite limit and may be represente«l by such a formula as that of hicks, namely a 2 y=a-r | | tut = mm where a, pz, a are constants calculated from the observed lincs and #2 takes successive integral values. a is the limit of the series, and r the rydberg constant. the canstituent series are of different types, having names and symbolical representations as fcllows:— principal series =1s—mp sharp series = 21-115 diiluse series =2p—ind fundamental series = 3«1— nef i g series, etc. =41 ig clc. in each case the first term denotes the limit of the series, while the second represents a sequence of ferns, in which mm has succes- sive intesral values. a group of such series constitules a sysicm, and there may be two or more systems in the same spec- trum, as the singlet and triplet systems of the alkaline earths. it will be observed that in a given system the limit of a series is a term of one of the other series, and that a line 1s always repre- sented as the difficrence of two terms. as first indicated by ritz in his “ combination principle,” other series may exist, such as 2s—mp, 3p—aid, etc. in a singlet system, each of the terms has but one value, and all the lines are single. ina doublet system, s has one value and all the rest two values. the principal series thus consists of doublets, 1s —ntp, and is—mpi, diminishing in separation towards the limit; the sharp scries of equidistant pairs ips —u:s, 1); —mis. the difuse doublet serics does not consist of double-dowolets, as might have been expected, but of groups of three, one possible component being ruled out by a principle of sclection to be inentioned later. in a triplet system, all but the s terms have three values, and certain sntercombinations with the associated singlet system also occur. origin of specira.—the theory of n. bohr’ offers a very sat- isfactory explanation of the spectra of hydrogen and enhanced helium, and gives a physical meaning to the spectroscopic terms in general which has proved very fruitful in suggesting new directions of research. according to this theory, the atom of 1 afanthly notices r.a.s., vol. 62, 63 (1913); phil. trans., a, vol. 214 (1914). 2 phil. mag., vol. 29, 284 (1918). 8 annalen d. phys., vol. 50 (19t6). 4 phil, mag., vol. 26, pp. 1-25; 476-502; 857-875 (1913); vol. 27, pp. 506-524 (1913); vol. 29, pp. 332-335 (1915). ‘623 an element consists of a positively charged nucleus with an appropriate system of electrons in orbital motion around it, such that the total negative charge of the electrons is equal to the positive charge of the nucleus. nearly the whole of the mass of the atom is concentrated in the nucleus, which is very small in comparison with the distances separating it from the electrons. when an electron is removed from its normal position by the application of an external stimulus, it may traverse temporarily one or another of certain orbits determined by quantum con- sldcrations and is said to be ‘ excited.” in each of these orbits it has a certain amount of energy, and there is no radiation so long as the electron revolves in such an orbit. when the electron returns to its normal position 11 comes to an orbit in which, for equilibrium, it must possess less energy than it had in the tem- porary orbit. the difierence of energy, which is preporticnal to the difference of the corresponding terms, is emitted as a homogencous radiation, and gives rise to a definite spectral line, while, if the electron occupies successively different orbits on its return, several lines will be produced in succession. the actual spectrum at any moment is the summation of the different lines vickled by atoms in different states. the differences between the are and enhanced spectra receive explanation on the bohr theery. the lines of an are spectrun. are supposed to be generated by the disturbance of a sacle outer clectron and its subsequent iateraction with the nucleus and remaining clectrons. when two electrors are re- moved from their normal pesitions, and one remains at a creat distance, the return of the second ciectron generates an entirely lifferent spectrum consisting of the enhanced imes. an atem which has lost an clectron is said to be “ienised ’: wit has lost two electrons it is doubly ionised, and so on. assuming the hydrogen atom to consist of a nucleus and a single electron, the energies of the possible orbits can be calcu- jated, and are found to be proportional to the observed terms rim? helium, the next lightest element to hydrogen, has two electrons, and the mathematical problem of determining their motion has not yet been solved. if one of the electrons is re- moved, however, the atom is similar to that of hydroccn, except that the nucleus has a double positive charge and four times the mass. the resulting enhanced terms are therefore calculable. they again have the form rein?, but r has now a much larger ralue than it has for hydrogen. it is represented in both cases a a snnple : ae. by the expression. == —-, = =-— = where en and 1,31 are te- ch® — at-e ni spectively the charze and mass of the electron and nucleus, & planck’s constant amd c is the velocity of light. in the case of hydrogen e=e, and when the experimental values of the varicus quantities are substituted in the formula, the series constant is re- produced with remarkable accuracy. the second factor inereases with m, so that it will be slightly greater for helium than for hydrogen. also, the double nuclear charge makes the first aelor in the expression four times as great for enhanced helium as it is for hydrozen. these theoretical requirements have been completely verilied ly experiment. fowler, calculating rk for hydrogen and enhanced helium from the observed lines, used the theoretical expressions to calculate the value of mjat—i.c., the ratio of the masses of the hydrogen atom and the clectren— and obtained a result in very close agreement with that arrived at by direct measurement. moreover, he has shown® that in the more complicated spectra of the alkaline earths the enhanced line terms are also represented by formulae in which r has four times its value for arc spectra. more recently, f. paschen has identified 4 r series and 9 r series in singly tonised and doubly tonised aluminium respec- tively,® while fowler has traced series of silicon for which the scrics constant has the successive values r, 4r, or and 16 r.? for convenience of reference these spectra are designated si 5 phil. trans., a, 214, p. 254 (1914). 6 annalen d. phys., vol. 71, p. 142, and p. 537 (1923). 7 phil, trans., a, vol. 225, p. i (1925). 624' i, si it, si (12, si iv, and similarly for other elements. these results for elements at successive stages of ionisation give further strong support to bohr’s theory by their verification of the changing value of the series constant as electrons are detached from the atom one by one. further developments of the theory, taking into account the varia- tion of the mass of the electron with velocity required by the theory of relativity, have indicated that the lines of the hydrogen and enhanced helium series are complex, and under high resolution should appear to consist of several components. this has been verified by paschen,! who found results for helium in remarkable agreement with the predictions of sommerfeld. the intensities of the several com- onents also are in the ratio calculated by sommerfeld by a special y pothesis. resonance and tonising potentials —-strong support for the bohr theory is also given by experiments in which atoms are bombarded by electrons of regulated speed. if an electron of charge e is made to fall through a potential difference v, it acquires a quantity of energy ev, expressed in appropriate units. if such an clectron bombards a neutral atom, tt is found that no change takes place until v reaches a certain value, when there is a sudden radiation of energy corresponding to a particular line—usually a strong flame line —in the spectrum of the bombarded atom. according to bohr’s theory, the energy in this radiation is equal to ky, and experiments with several elements show that this critical value of vis determined by the relation ev —kv. this means that the energy of bombardment has been just sufficient to remove the outermost electron in the atom from its normal orbit to the next permissible orbit, and, on its return, the electron restores the energy in the form of monochromatic radia- tion of the appropriate frequency. if v is expressed in volts, the wave-number of the emitted line is numerically equal to v x8t102. the value of v when emission first takes place is known as the “resonance ”’ or “ radiation ’’ potential. further, by increasing yv, it is possible to remove an electron from the atom altogether. if the energy ev’ in this case is equated to fy, the resulting value of » is found to be equal to the largest term in the spectrum. this term would correspond to the innermost orbit of the electron, and the result indicates that the energy which must be applied to remove an electron from the atom is just equal to the energy possessed by the electron when revolving in its normal position. the potential re- quired to remove an electron from the atom is known as the “ lonisa- tion potential.’” it has been determined experimentally for a number of elements, and has been found to be in complete agreement with the orbital energy as calculated from the largest term in the spectrum. the bohr theory thus presents a simple picture of the processes taking place in expcriments of this type. zeeman and stark efects.—the resolution of spectral lines under the influence of intense magnetic fields—usually known as the zeeman effect—has been extensively studied for a large number of elements. somewhat similar effects, but much greater in magnitude, produced by an electric ficld, have been brought to light by stark and examined in considerable detail by a number of workers. j. w. nicholson and t. r. merton? have shown that the stark effect may operate to an appreciable extent in an ordinary vacuum-tube dis- charge, causing a broadening of the lines. both the zeeman and stark effects have been treated on the basis of the bohr theory with success. a comprehensive set of rules for the calculation of the zeeman effect for any combination has been given by lande. the more complex spectra.—the spectra of the higher groups of elements are much more complicated than those of the earlier groups, and few series of the same simple types have been iden- tified. an advance of great importance, however, was made in professor fowler’s laboratory by m. a. catalan‘ in the identi- fication of complicated groups of lines in the spectrum of man- ganese and other elements which could only be accounted for on the supposition that spectroscopic terms may have a greater multiplicity than that represented by triplets. the groups of lines in question, showing internal regularities, were named multiplets, but it may be noted that the pd and df combinations in a triplet system, each consisting of six lines, are also entitled to the same designation. with this beginning, the analysis of complex spectra has proceeded vigorously, and the structure of even so compl cated a spectrum as that of iron has been eluci- dated. thus, quartet, quintet, sextet and still higher systems of spectroscopic terms are now recognised. theoretical work based on quantum considerations has also progressed rapidly, and it is now possible to state quite briefly 1 annalen d. phys., vol. 50, pp. 901-940 (1916). 2 phil. trans., a, vol. 216, p. 459 (1916). 3 zeit. f. phys., vol. 15, p. 192 (1923). 4 phil. trans., a, vol. 223, p. 127 (1922). spectroscopy the rules which have been deduced from the analysis of spectra. following sommerfeld and a. lande,5 a spectroscopic term in general may be represented by 2';;, where 7 is the “ principal ” quantum number (related to m of the previous formulae), & is the “ rotational ” or “ azimuthal ” quantum number, j a char- acteristic “inner” quantum number, and ¢ represents the maxi- mum multiplicity in the system to which the term belongs. it is convenient in practice to replace x, by ms, mp, etc., as used previously, it being understood that k=1, 2, 3, ... for terms of s, p, d .. . types respectively. the various components of a multiple term are then distinguished by using the relevant values of 7 as subscripts, which differ in this respect from the subscripts adopted until quite recently. when necessary to avoid ambiguity as to the system of terms under consideration, the multiplicity is indicated by an index figure on the upper left side of the term symbol. thus 3%. indicates the term 3p of a triplet system which has the inner quantum number 2, it being understood that all » terms have k=2. the characteristics of the first six groups of terms are shown in the appended table. doublet n lf so singlet r=i r=2 k=2 p, pi pz k=3 i). ds dy k= | f, f; {3 k=1 s| triplet sa quartet c= r=4 k=2 py pr pe pr pz ps k=2 a; ds ca; dy de ds d, k=4 f., {; f4 f. fs fs f. k=1 ss. quintet sx sextet r=4 r=6 kes pi pz ps pz ps pa k=3 dy ch che ds day d, de ds dy ds k=4 f; fy fs fi, fs f, t» fy fy fs fs the combinations which ordinarily occur within the same sys- tem are those for which & changes by one unit; i.e, ak= 1, and the p terms, for example, combine with s and d but not with f. the combinations are further restricted by the condition that aj= +1 or o, with the exclusion of the transition o to 0; thus, in a triplet pd combination the nine lines which are arith- metically possible are reduced to six, namely, pods, pods, pods, pide, pds and pod;. when two or more systems occur in the same spectrum, the &and7 rules are still applicable, and intercombina- tions are restricted to those fur which a,s= +2. thus, in the alkaline earths there are both singlet and triplet systems, but in the combination of 's» with *9, the application of the 7 rule shows that the only possible line will be '.so—*1. as an illustration of one of the simpler multiplets, the following quintet pd group in the arc spectrum of iron may be quoted, the positions of the spectral lines being indicated by their wave-numbers, and the relative intensities by figures in brackets following the wave-numbers:— do 1 d» ds dy 36431 (10) 9036521 (20) 184 36705(15) pr 22 booz 36269(20) 184 36453 (30) 288 30741 (40) pe 288 36351(25) 416 36767 (60) ps anomalous terms.— the terms of ordinary types which have so far been considered are regarded as originating in the transitions of a single electron between different virtual orbits. it has been found, however, that in many spectra there are other terms resembling these in regard to their inner quantum numbers and zeeman elfects, but differing from them in following the combination rule ak =o; these are distinguished by writing them in the form s’, p’, d’, etc., and are called ‘* primed terms.’”” a whole series of p’ terms has been found in the alkaline earth spectra by h. n. russell and f. a. saunders® who have shown that the limit of the series is greater than that of any of the regular series. in other words, in some of the combinations pp’ the atom radiates a greater amount of energy than that due to the largest jump which a single electron can make. the adopted explanation is that two of the outer electrons are dis- placed, and that the changes of energy of both are united in the emission of a single quantuin. 5 zeit. f. phys., vol. 15, p. 191 (1923). 6 -lstrophys. jour., vol. 61, p. 38 (1925). 36063 spectrum ee other anomalous terms are similar in their combining propertics to terms of regular types from which they differ only in the fact that they have no place in the rydberg series. these are <listin- guished by doubly accented iectters, as p”, d”, etc. spectra and the periodic table.—attention is being drawn more and more to the relation of spectra to the periodic table of the elements. it has long been known that, when doublets or triplets occur in the spectra, the wave-number scparations of their components (which are constant in the sharp and diifuse series) are approximately proportional to the squares of the atomic weights of the elements producing the spectra—so long as those elements belong to the same family sub-group. the zeeman effect also is generally the same for lines of correspond- ing series in the spectra of clements belonging to the same group. but perhaps the most comprehensive connection of spectra with the periodic table is established by the ‘ alternation law ” and by the “‘ displacement law ” of kossel and sommerfeld. the spectra of a large number of elements have now been successfully analysed, and it has been found that there is an alternation of even and odd multiplicities in successive columns of the periodic table. this has been most completely verified for the fourth period of the elements, for which the following details have been given by w. if. meggers:— calan: 1 ju lurliv! v i vil virlvir element k-19 va-20/8c-21|ti-22] ¥-23 |cr-24|ma-25 |pe-26|c0-27)ni-28 systems i i ts 2 2 2 2 3 3 3 3 3 wn wa na 5 en a 4 nn + “j “i the numbers following the chemical symbols are the atomic numbers of the elements, and those below represent the multi- plicities of the systems of terms which have been found to com- bine in the production of the actual spectra. elements in the same column of the table have similar multiplicities. the displacement law states that, when an element is ionised, the enhanced series take on the same type of complexity as the arc series produced by the element to the left (7.e., in the pre- ceding column) in the periodic table. ‘uhis has been verified for a considerable number of elements and 1s probably true for all. the same rule applies also to the spectrum of an element at suc- cessively higher stages of ionisation. as a result mg if, al iil and si iv, for example, all yield spectra of the same type as that of na i, only one valence electron being left in each case; these spectra only differ in the displacement of corresponding lines to shorter wave-lengths and in the wider separation of corresponding pairs as the degree of ionisation is greater. con- tinued observations of this kind may be expected to contribute largely to our knowledge of the structure of atoms. suggestion as to the arrangement of clectronic orbits in a large number of elements, based upon spectroscopic data, have been made by n. bohr,” and important modifications of bohr’s scheme have been proposed by e. c. stoner.’ band spectra.—a typical band consists of a large number of faint lines, which become especially crowded towards the ‘‘ head,” where the intensity appears greatest. groups of such bands frequently occur, partially overlapping, and there may be several groups in the same spectrum, so that a band spectrum may include thousands of component lines. in each band it is usually possible to pick out a number of regular sequences of lines, each of which may be approxi- mately represented by a deslandres' formula v=a+bm+cne, where a, b, c are constants calculated from the lines and 2 is the number of a line counted from a suitable starting point. the heads of the bands may be represented by a formula of the same type. the theory of band spectra, on quantum principles, has lately been greatly advanced. it is supposed that in a molecule the total energy is the summation of the energics due to rotation of the mole- cule as a whole, to vibrations of the atomic nuclei and to move. 1 proc. nat. acad. sci., washington, jan. (1925). 2ann. d. phys., vol. 71, 260 (1923). 3 phil. mag., vol. 48, 719 (1924). ments of the electrons within the molecule. hf a change occurs in one or more of these, the total energy will be changed and, exactly as in line spectra, the difference will be radiated as a single quantum the complexity of a band spectrum is thus due to the great variety of possible states of the molecules by which it is emitted (see som- merfeld, op. c7., chap. vii.). the solar spectrum.—aa striking feature of continued work on the solar spectrum is the identification of a large number of faint lines with lines composing the bands of certain compounds, in addition to the band lines of carbon and cyanogen previously recognised by rowland and lockyer. the peculiarities of the region about the g group of fraunhofer have been shown by h. f. newall? to be due to the absorption of the well-known hydrocarbon band a 4315, and the group p has been found by a. fowler and c.l.l. gregory® to include the strong ultra-violet band of ammonia having its maximum near \ 3360. in addition, the band of luminous water vapour beginning at xx 3064 has been found by fowler® to be present in the solar spectrum. a large number of previously unknown solar lines have thus been accounted for, and it is not improbable that the thousands of faint lines which remain unidentified may eventually be traced to other band spectra. stellar spectroscopy—our detailed knowledge of the spectra of the stars has been greatly advanced by the use of the large telescopes which have been erected, and considerable progress has also been made in the interpretation of the stellar lines through experiments in the laboratory. in particular, the use of stronger discharges than had previously been employed has led to the discovery of new lines of several clements, which have been identified with lines occurring in the hotter stars. certain lines of the wolf-rayet stars, for example, have thus been traced to carbon by t. r. merton,’ and others to oxygen by a. fowler and j. brooksbank.* the general outcome of the experi- mental reproduction of stellar lines is to support the view that the order in which the different classes of stars had been arranged is a true temperature sequence. this order, previously indi- cated by secchi and vogel, is now generally expressed by the classification introduced at harvard by e. c. pickering, in which the most important classes, passing from the white to the redder stars, are designated by the letters b, a, f, g, k, m." on passing from the relatively cool m stars to the hot b stars, it is necessary, in accordance with the work of lockver, to employ a gradually increasing stimulus in order to excite the spectra which appear at successive stages of the stellar sequence. the theory of thermal ionisation has been further developed by r. h. fowler and e. a. milne! with important results. bripliography.—to the works mentioned in the earlier article, the following shoukl be added: j. m. eder and e. vatenta, alas typischer sperktren (igi); j. stark, die atoniionen chemischer elemente und ihre kanalstrahienspecira (1913); p. zeeman, re- searches tn magneto-optics (1913); t. lyman, the spectroscopy of the extreme ciltra-violet (1914); p. d. foote and f. l. mohler, phe origin of spectra (1922); a. fowler, series in line spectra (1922); w. m. hicks, the analysis of spectra (1922); h. holst and hh. a. kramers, fhe :ltom and the bohr theory of its structure (1923); <a. sommerfeld, atomic structure and spectral lines (1923); e. back and a. lande, zeemaneffekt und multiplettsirukiur der spektrallinien (1925); miss c. hl. payne, stellar atmospheres, harvard observatory monographs, no. 1 (1925). (a. f.) spectrum.—the word spectrum was used by newton in 1672 to indicate the elongated array of coloured lights, which appears when a beam of sunlight is passed through a triangular 4 monthly notices, r.a.s., vol. 76, 640 (1916). ° phil, trans., a, vol. 218, 351 (1917). 6 proc. roy. soc., a, vol. 94, 472 (1918). * proc. roy. sec., a, vol. 91, 498 (1915). 8 monthly vol., r.a.s., vol. 77, 511 (1917). °the work of hl. n. russell, in general agreement with that of lockyer, renders it probable that the true sequence is from m to 1 with increasing temperature, and thence from b to m with decreas- ing temperature, the density increasing throughout. stars of rising temperature, on account of their great volume, have been called ‘“ giants,” those of falling temperature ‘‘ dwarfs.” differences be- tween the spectra of giants and dwarfs of the same spectral class have been found by adams (see monthly notices r.a.s., vol. 81, p. 334) (1921). 0 proc. phys. soc. bond., vol. 36 (feb. 1924). 626 glass prism and, after refraction and dispersion, is received on a distant white screen. the name is firmly established for the orderly array of col- oured lines, each corresponding to a single component, seen when a sample of light to be analysed has passed through a very narrow slit of a suitable spectroscope. ‘“‘ spectrum,” is used in an extended sense to cover the whole rarge of rays studied by methods capable of detecting not only visible rays but also ultra- red and ultra-violet rays, wireless waves, x-rays and ‘y-rays (see rays). ulira-violet rays——the existence of ultra-violet rays was discovered in 1801 by ritter, who found that silver chloride was blackened by rays more refrangible than the utmost limits of the visible rays of the spectrum. the ultra-violct rays were examined in detail by stokes in 1852 by the use of fluorescent substances to which attention had been called by sir john ier- schel in 1845 in researches on what he termed the “ epipolic dis- persion of light ” by solutions of quinine. stokes found that quartz prisms and lenses were far more transparent to ultra- violet light than glass, and that they revealed the existence of an invisible region extending as far beyond that previously known as the latter extends beyond the visible spectrum, and exhibiting a continuation of the fraunhofer lines. he obtained evidence that the limit of the solar spectrum in the more re- frangible direction had been reached; for the arrangement which revealed by means of fluorescence the existence of what were evidently rays of higher refrangibility coming from electric sparks failed to show anything of the kind when applied to the solar spectrum. measurement of wavelengths -—on the undulatory theory of light, wavelengths can be assigned to rays of every degree of refrangibility by the use of diffraction gratings. a grating is a plate of glass or of speculum metal worked plane (or concave) and highly polished, and on it fine straight parallel lines cre ruled very close together at very regular minute intervals. thomas young in 1802 describes the use of a grating ruled on glass with 500 lines to the inch; and, in his developments of the undulatory theory and of the principle of interference, he de- duced rough values of the average wavelengths of the various colours. modern gratings are ruled with 10,000 to 20,000 lines to the inch. they serve admirably for optical work of high pre- cision. but when we have to deal with very short wavelengths such as have been discovered in the last 30 years, a spacing of lines comparable in closeness with the spacing of atoms in solid substances is required. angstrem unit.—waveclengths are measured in terms of a unit known as an angstrem unit (a.u.) or tenth-metre. such i i ; a unit is th of a metre = — in cm. the vis- 10 10,000 000,000 108 ible parts of the spectrum extend from about 3,900 a.u. to 7,600 a.u., nearly an octave in the musical sense of the word. in terms of these units, stokes would seem to have seen lines as far as 1,900 a.u. in the spectra of electric sparks between terminals of metallic lead. with fluorite prisms and lenses (in rarefied hydrogen to avoid the absorption due to air) schumann photographed spectra extending to 1,000 a.u.; lyman has reached 600 a.u.; and millikan has penetrated as far as 200 angstrem units. | | ultra-red rays.—rays of invisible heat were discovered by william herschel in 1801 in the ultra-red region.1. langley in 1886 traced, by the use of the bolometer, rays of radiant heat as far as wavelength 182,000 a.u. and rubens penetrated as far as 610,000 angstrem units. thus our knowledge of the spectrum extends from 200 to 610,000 a.u., of which only the region between 3,900 and 7,600 a.u. is visible to the eye as light. (see rays.) maxwell’s theory.—in 1861-5 maxwell developed, out of the physical conceptions of faraday, his dynamical theory of the electromagnetic field. he gave reasons for believing that any change in the electrification of a conductor or in the magneti- 1this region is very frequently called the infra-red region, and its rays the inira-red rays (see rays). : spectrum sation of a body is accompanied by the propagation of corre- sponding disturbances through the surrounding medium. his equations showed that the rate of propagation should be ex- pressed by the same number which expresses the number of electrostatic units in one electromagnetic unit. the comparison of available measures of these electrical quantities gave a number corresponding with the velocity of light. maxwell adds:— at the outset of this paper, the dynamical theory of the electro- magnetic field borrowed from the undulatory theory of light the use of its luminiferous medium. it now restores the medium after having tested its power of transmitting undulations and the char- acter of those undulations, and certihes that the vibrations are transverse and that the velocity ts that of light. what then is light according to the electromagnetic theory? it consists of alternate and opposite rapidly recurring transverse magnetic disturbances accom- panied with electric displacements, the clirection of the electric dis- placement being at right angles to the magnetic disturbance, and both at right angles to the direction of the ray. ledge and hertz—in 1887-8 oliver lodge and hertz were, each unknown to the other, running a neck-to-neck race, to crown maxwell’s achievement by devising experimental meth- ods of producing and detecting electro-magnetic waves of very high frequency—about 100,000,000 per sec.—very high in comparison with those in common use in ordinary electrical experiments, but very low in comparison with those attained in langley’s examination of the ultra-red parts of the prismatic spectrum. hertz succeeded in measuring, within the walls of his laboratory, the wavelengths of the periodic disturbances emitted by the spark-discharge of an electric condenser. and from these pionecring experiments the whole system of wireless telegraphy (g.z7.) has been evolved by the inventive genius of many workers. the wavelengths of hertzian or wireless waves range from a few hundred metres to long waves of about five km. and more. a -rays (g.v.).—the x-rays discovered by rentgen in 1895 are now known to be identical in nature with light but to be of exccedingly small wavelengths. a. laue in r9or2 made a funda- mental discovery when he passed x-rays through crystals and found evidence which at one and the same time disclosed the orderly atomic siructure of such bodies and the undulatory nature of x-rays. w. l. bragg in 1913 discovered that x-rays could be reflected by the cleavage planes of crystals at nearly glancing incidence; and he and his father sir w. h. bragg de- vised the x-ray spectrometer, by means of which they measured the wavelengths of specially selected x-rays and studied the structure of crystals. their methods were applicable both to the electrical detection of the rays and to photographic records of the “ spectral lines ”’ of such characteristic rays of the elements as were discovered by barkla in 1904. the tactical array of the atoms in a crystal serves the same sort of purpose as the sys- tematic spacing of lines on an optical grating and wavelengths as smal] as 1 a.u. were measured (see crystallography). mosely in 1913 discovered that in the high frequency spectra of the clements a wonderful progressive increase of wavelengths of homologous lines was disclosed by the elements when arranged in order of atomic weight (see atomic wficnts); and theoretical considerations led him in 1914 to prove by measurement of photographic records that (1) every element from aluminium to gold is characterised by an integer which determines its x-ray spectrum, and (2) this integer n, the “ atomic number ” of the element is identified with the number of positive units of elec- tricity contained in the atomic nucleus (see isotopes). he thus corroborated the conceptions of rutherford (1911) and of van den broek (1913) of the nature of this constituent of an atom, and, besides establishing a fundamental principle in atomic physics, provided measures of wavelengths in the ex- tremest regions of the ultra-violet spectrum. x-rays used in physics and medicine range from about o-or a.u. to 12 angstrem units. the highly penetrating rays are of short wavelength; the “ softness ” of rays increases with the wavelength. the y-rays emitted by radioactive substances are highly penetrating rays and have wavelengths less than 1-5 angstrem units. the recent observations of millikan in 1925 on spele—sperry a very highly penetrating radiation point to the existence of x-rays of wave-lengths even smaller than those of y-rays. thus our knowledge of the spectrum of clectromagnetic ra- diations may be said to extend over a range of more than 7o octaves, or from wavelengths from o-o01 a.u. to 10,000 km.; and within this range the visible rays occupy hardly an octave; the british broadcasting company utilise rays ranging over about 2 octaves; and only about 5 octaves in the whole range have escaped actual detection (see wireless). see e. c. c. baly, spectroscopy (1924). (h. f. n.) spee, maximilian, count von (1861-10914), german sailor, was born in copenhagen june 22 1861. he entered the german navy in 1878 and from 1887 to 1888 was itafenkomman- dant (commander of the port) in german cameroon. in 1908 he was made chief of staff of the german north sea command, and at the end of 1912 took over the far eastern squadron. on nov. 1 1914 he was engaged by admiral cradock off coronel, on the chilean coast, and succeeded in defeating the british squadron. on dec. 8 1014, however, he was attacked by admiral sturdee’s cruiser squadron. the germans were heavily defeated in the battle that taok place. this is usually known as the battle of the falkland islands, and in it spee himself went down with his flagship, the “ scharnhorst.’? (see coronel; falkland islands). speed indicator.—instruments used for meastiring either linear or angular speed are classed under the general heading of speed indicators. those used for indicating linear speed are used most extensively on motor vehicles and are known specifical- ly as speedometers, while instruments based on the same working principles but used to measure angular speed are known as tachometers. tachometers are always graduated in revolutions per minute, while speedometers are graduated in miles (or kilo- metres) per hour. the speedometer is driven from one of the road wheels or from the transmission tail shaft of the vehicle through a flexible shaft and a gear mechanism, and as its indications depend solcly upon the speed at which its own shaft revolves, account must be taken in calibrating it, of the diameter of the road wheel from which it is driven and of the gear ratio between this wheel and the speedometer shaft. inversely, if the scale is fixed, the gear ratio must be varied to suit the wheel diameter. odometers.—spceedomcters used on automobiles carry, in addition to the actual speedometer mechanism, a so-called odom- eter, which indicates the distunce covered by the car. this instrument comprises a number of dial wheels with numerals imprinted on their circumference, which show through a window in the face plate. the dial wheels do not move continuously but periodically, in steps of one-tenth revolution. usually there are two sets of such dial wheels, one set showing the total mileage covered by the car since the instrument was installed, the other the mileage covered on a single trip, the dials of the latter being returnable to the zero position at the end of each trip by means of a reset device. an odometer, therefore, may be re- garded as an integrator of linear motion; a corresponding instru- ment for integrating angular motion is known as a revolution counter and is much used on large, low-speed engines, in mechani- cal tests, and for many other purposes. | _ main principles there are three general principles on which the operation of speed indicators may be based, of which that of centrifugal force has been employed longest. with this, a pair of weights is carried on a revolving shaft, in such a manner that they may move out from the axis of the shaft under the influence of centrifugal force, while being forced toward the axis of the shaft by a spring, the motion taking place around fixed pivots. the centrifugal force on the weights increases with the speed, and the faster the shaft revolves, the farther the weights will move out from the axis of rotation. an indicating hand con- nected mechanically to the centrifugal weights may be made to indicate the speed of rotation of the shaft, or the speed of motion of a vehicle to one of whose road wheels the shaft is connected. the second principle very extensively used is that of magnetic drag. a cup of sheet aluminium is mounted on a spindle and 027 held in the “ zero” position by a spiral spring. a permanent magnet whose lines of force pass through the wall of the cup is rotated inside th®latter at a speed proportional to that to be measured. the rotation of the magnet induces currents in the aluminium cup and the reaction between these currents and the magnetism of the magnet produces a drag on the cup, which causes it to turn around its axis against the torsion of the spring, in proportion to the speed of the magnet. a scale printed on the outside of the cup shows through an opening in the face of the instrument, and the scale reading in line with a mark on the face indicates the speed. the third principle made use of is based on the fact that speed is the quotient of distance by time. instruments based on this principle comprise an odometer and a clockwork, measuring dis- tance and time respectively, and they effect the operation of division mechanically. (p. m. hh.) spender, john alfred (1862- ), british journalist, was born at bath, the son of a doctor, and educated at bath col- lege, and at balliol college, oxford, under jowett. adopting the career of a journalist, he was editor of the fastern morning news from 1886 to 1890, after which he came to london and was on the staff of the pall mail gazette. when the westmizstcr guszelte was established in 1893, he became an assistant editor and from 1896 to 1922 was its editor, leaving just after it became a morning paper. as editor of the westminster mr. spender won general respect, not only as a brilliant writer but as the possessor of an exceptionally fair and balanced mind. ie won, too, a high place in the councils of the liberal party, and his services to the state included membership of the royal commission. on divorce, and of the milner mission to egypt. a slight volume, the com- ments of bagshot (1914), reveals mr. spender at his best. he has also written zhe indian scene (1912); the foundations of british ‘policy (1917); and the life of sir. h. campbell-bannerman (1923). ils brother, harold, also known as a liberal journalist, and the author of a number of books. of travel and _ politics, died april 15 1926. spengler, cswald (1880-— ), german philosopher, was born at blankenburg in the harz may 29 1880. he studied mathematics and natural history at various german universities, and at the same time interested himself deeply in history and art. this early and rare combination of scientific knowledge with interest in the humanities is the foundation of the peculiar character of spengler’s work, in which unexpected parallels be- tween scientific truths of physics and mathematics and the artis- tic and other cultural achievements of an epoch of history are made clear. spengler’s thesis for his doctor’s degree dealt with “the metaphysical idea of the philosophy of heraclitus ” (iialle, 1904). from 1907 to rg11 he taught mathematics and physics at a school in hamburg. in 1911 he conceived the main theme of his work. abandoning teaching, he completed in 1914 the first version of der untergang des abendlandes, which re- mained unpublished until 1918. arevised and much enlarged edition was completed in 1922. an authorised translation of the first volume into english (fhe decline of the west), by charles francis atkinson, was published in 1926. the work has had great popularity and iniluence. > | other works of spengler, chiefly concerned with contemporary political problems are: preussentum und sositlismus (1920); pessim- ismus (1921); politische pflichten der deutschen jugend (1924); neubau des deuischen reiches (1924). see m. schroeter, der streib win spengler, etc. (1922). sperry, elmer ambrose (1860- ), american inventor, was born at cortland, n.y., oct. 12 1860. his education was received at cortland normal school, 1876-9, and cornell uni- versity, 1879-80. he founded the sperry electric co., chicago, in 1880, engaging in the manufacture of arc lamps, dynamos and other electrical appliances, and also the sperry electric co., cleveland, o., manufacturers of electric cars. this last-named company was sold to the general electric co., new york city, in 1894. he perfected one of the first arc lamps, and invented aero- plane and ship stabilisers, the highest intensity searchlight (1,500 million candle power), and many special devices for the u.s. navy. his most important invention was the gyro-compass 628 (patented in 1918), designed to obviate magnetic disturbances incident particularly to armoured ships and submarines. he further devised the gyro-pilot, a device for th@automatic steering of a ship. in 1910 he became president of the sperry gyroscope co., brooklyn, n.y., a company formed to manufacture his in- ventions. in 1915 he became a member of the naval consulting board and chairman of its mines, torpedoes and aeronautics committees. in 1883 he erected the highest electric beacon in the world on lake michigan (350 feet). spielhagen, friedrich von (1829-1911), german novel- ist (see 25.667), died at charlottenburg, berlin, feb. 25 rort.