Feb. 28, 1884] 



NA TURE 



421 



it is desired to determine the true curve uf diurnal variation of 

 the wind velutity at any given ttation. In ihe ca^e of proceed- 

 ing by the ordinary routine of hourly sums and mean?, it will 

 le found that the occurrence of a hi<;h wind or gale on a single 

 day will vitiate the results for a considerable period of time. 



If, on the olher hand, instead of doing this, a drawing or 

 photograph be made on one sheet of ihe daily curves for a few 

 weeks, it will be found that the traces for the days free from 

 storms will lie so fairly close ti'gether or upon one another, that 

 little ditBculty will be found in selecting or drawing through 

 them a curve representing the general run of the group. Several 

 sets of curves having been so treated, the typical curves must be 

 in turn themselves superimposed, and through them another 

 curve drawn, which will be still less affected by abnormal 

 movements ; s ) eventually the true curve of diurnal variation 

 would be arrived at. 



In the case of subjecting photographic traces, c".^. barograms, 

 thermograms, electrograms, magnetograms, &c., to this treat- 

 ment, it would be advisable to employ secondary impressions or 

 prints from the original curves, in order that the composite 

 produced might consist of dark lines on a white bacl<ground ; 

 not the reverse, which would be comparatively useless for the 

 purpose. 



For the reduction of anemograms, rain, and sunshine curves 

 by this method, it will be necessary lo make drawings or tracings 

 first from the curves, giving the hourly values separated, as is 

 done in the diagrams published in the Quarterly Weather Reports 

 of the Meteorological Office and in the Kew Times curves. 



Another application of the method of composite drawing will 

 serve to facilitate the acquisition of a knowledge of the general 

 distribution of weather systems over large tracts of the earth's 

 surface. To do this, a series of weather charts should be taken, 

 and selecting certain prominent features, such as the centres of 

 cyclonic and anticyclonic disturbances, day by day their positions 

 should be marked off upon one chart. This being done in a 

 sufficient number of cases and combined, a repetiiion of the 

 process would enable a determination to be made of the average 

 distribution of these systems for a given season. 



The author illustrated his proposed applicati')ns of the method 

 of composite portraiture by three examples, which were exhibited 

 lo the meeting of the Society. The data treated in every ca^e 

 were chosen at random, and therefore may be considered as 

 indicating the applicability of the process to meteorological work 

 in general. 



In the first example the mean diurnal variation in the wind 

 velocity at the Kew Observatory, Richmond, was determined for 

 three months — August to October, 1S79. Taking the hourly 

 values of the rate at which the wind was blowing from the 

 Meteorological Office publications, they were plotted down on a 

 conveniently open scale, a fortnight's superimposed curves being 

 on a sheet. Through the fourteen curves so drawn in pencil a 

 mean curve was traced in red. This roughly represented the 

 average daily variation during the fortnight. 



The pair of fourteen-day curves being superimposed ou a third 

 sheet, a third trace drawn between them was assumed to be the 

 mean trace for the month, and finally combining the three so 

 derived months' traces, it became easy lo draw the final curve 

 showing the mean diurnal variation of wind velocity during the 

 quarter in question.' 



The second experiment was an attempt to obtain a monthly 

 mean of the barometer directly by the graphic method. Taking 

 advantage of a self-registering aneroid being on trial, its traces 

 were utilised for the month January 8 to February 7, 1883. 

 These were copied off on a sheet of tracing paper, ruled jo as to 

 comprise one day's curve only. The tracing paper was then 

 folded vertically, so as to compress the curves, and the mean 

 positions of the traces were drawn on the folds. After four 

 foldings a point was readily fixed upon as the position of the mean 

 of the month, and the value of this point referred to the scale of 

 the instrument. The resulting value f^r the mean barometric 

 pressure of the month very satisfactorily agreed with the value 

 defermined by calculation from the barometer readings taken 

 daily at the Observatory. 



The third series of illustrations represented the general posi- 

 tions of the centres and the contours of the areas of maximum 

 ' It must be remarked that a due proportion should be preserved between 

 the scales of the ordinates and abscissEe, for unless this is done the combined 

 traces may appear merely as a mass of confused lines. Such was the case 

 in some experiments made by the author, when he attempted to derive 

 mean curves directly from the zinc templates engraved at the Meteorological 

 OfSce for the Quarterly Weather Reports, kindly placed at his disposal by 

 Mr. Scott. 



and minimum barometric pressure over the Atlantic during 

 January, February, and March, 18S1. A number of blank charts 

 were worked off by the chromograph, on tracing paper, to the 

 scale of the international synchronous charts of the U.S. War 

 Department Signal Service. Tracings were made on one sheet 

 m blue pencil of the cyclonic centre for each day of the month, 

 aiid then on another a similar set of tracings in red of the anti- 

 cyclonic centres. Having from these drawn the prevailing posi- 

 tions and areas of the systems for the month, it was easy to draw 

 another chart with the general distribution for the quarter. 

 The diagrams were seen, on comparison, to differ materially 

 from those drawn for the monthly means of the observations. 

 In suggesting the composite method of treatment of meteoro- 

 logical data, the author is fully aware that a somewhat similar 

 process has been already applied in the determination of the 

 ladiant points of shooting stars, and would also desire to state 

 that the process is not by him considered as equalling or even 

 approximating in accuracy that of employing the harmonic 

 analyser in computing the periodical variations of the elements. 

 As, however, that instrument is not at the command of many 

 investigators, he is of opinion that the labour of reduction may 

 in many caies be saved by making use of the graphic or com- 

 posite, instead of the purely numerical, method. 



UNIVERSITY AND EDUCATIONAL 

 INTELLIGENCE 



Cambridge. — Dr. Hans Gadow, Strickland Curator, has been 

 approved as a Teacher of Comparative Anatomy ; Mr. L. 

 Humphry, M. B. , as a Teacher of Pathology ; and Mr. F. H. 

 Neville as a Teacher of Practical Chemi>try. 



Messrs. J. \V. Hicks, K. D. Roberts, and A. S. Lea are ap- 

 pointed Examiners in Natural Science in the Special Examina- 

 tions for the ordinary B.A. degree. 



The Examiners' Report on the Special Examinations in Natu- 

 ral Science states that there was no improvement in the book 

 w ork, but the practical work was more intelligeirtly done. The 

 few candidates in Geology did well. Botany was ill done. In 

 Zoology the candidates did well. 



Mr. J. A. Lyon (Clare College) has been appointed to the 

 new office of Superintendent of the Mechanical Workshops. 



SCIENTIFIC SERIALS 



Bul'etin de I' Academic Royale de Belgiijue, November 3, 1883. 

 On the anatomy and histology of a new specie^ of derostoma 

 {D. benedcitu), by M. Francotte. — Report om the work still 

 required to complete the geodesic survey of Belgium, by Capt. 

 I-'elporte. — Observations on the periodic shooting stars made at 

 Louvain in 1S82-83, by M. Terby. — Influence of magnetic dis- 

 turbances on the scintillation of the stars, by M. Charles 

 Montigny. The paper is accompanied by various comparative 

 tables showing the intensity of scintillation before, durin,', and 

 after the magnetic disturbances in dry and wet weather. — 

 Summary report on the researches undertaken at the Ostend 

 biological station during the summer of 1883, by Edouard van 

 Beneden. Amongst the remarkable objects fished up near this 

 station were a torpedo of unusual size { Torpedo marmorata)^ a 

 fine specimen of Labrus maeiilalus, an Amphioxus lanceolaius, 

 and an unknown species of Scopelidx, referred by GUiiiLer of 

 the British Museum to the Odontostomus, or some albcd genus. 

 — On the observation of very rapid movements, especirlly when 

 occurring periodically, by M. J. Plateau. — Analytical study of 

 the volcanic ashes which fell at Batavia during the eruption of 

 Krakatoa on August 57, 1883, by M. Renard. The author 

 concludes that these ashes are formed by the pulverisation of a 

 fluid igneous mass, whose particles, projected by the expansion 

 of the gases, are subjected to rapid cooling during their passage 

 through the atmosphere. Nothing was detected to indicate the 

 direct action of vapour of water in volcanic disturbances. — On 

 the perfect elasticity of solid bodies chemic.illy defined. New 

 analogy between solids, fluids, and gases, by W. Spring. Here 

 are embodied some of the results of the researches conducted by the 

 author for several years on the action of pressure on solids reduced 

 to a powder. The main object of these researches was to ascertain 

 by experiment whether it be possible by means of pressure perma- 

 nently to diminish the volume occupied by a given weight of a 

 solid body chemically defined. As a general result, a slight mcrease 

 of density w a^ obtained under a pressure of 20,000 atmospheres. 

 But, this once realised, most bodies resisted all further perma- 



