On the Figures obtained by streuing Sand on Vibrating Suifaces , commonly called Acoustic Figures .
By CHARLES WHEATSTONE , Esq. Communicated by MICHAEL FARADAY , Esq. , D.C.L. F.R.S. , Ssc .
Sc. 8sc .
Received February 12 , -Read March 14 , 1833 .
§ 1 .
tHALF a century has nearly elapsed since the attention of philosophers was first called to the curious phenomena exhibited when sand is strewed on vibrating surfaces .
Long before this time , GALILEO had noticed that small pieces of bristle laid on the sounding-board of a nlusical instrument , were violently agitated on some parts of the surface , whilst on other parts they did not appear to move ; and our own countryman Dr. HOOKE , whose sagacity in anticipating many of the discoveries of later times has been so frequently remarked , had proposed to observe the vibrations of a bell by strewing flour upon it .
But to CHLADNI is d-ue the sole merit of having discovered the syimmetrical figures exhibited on plates of regular forms when caused to sound .
His first investigations on this subject , Entdeckungen iiber die Theory des Klanges , were published in 1787 ; this work was followed by his Akustik in 1802 , and his Never Beytrage sir Akustik , 1817 .
A French translation , by himself , of his second work was published at Paris in 1809 .
All the figures obtained by CHLADNI on square surfaces are delineated in pages 611 , 613 , 615 ; they are copied from the Never Beytraige , which woLrk contains his most mature experiments ; but not having been translated eitherinto French or English , it is but little known in this country .
The following are the general results deduced by CHLADNI from his observations respecting these figures : his works may be referred to for the details omitted , and for those concerning the vilbrations of plates in general .
In all the modes of vibration of a square or rectangular plate , the figures , even if they consist of diagonal or tortuous lines , may all be referred to a certain number of nodal lines in the two directions parallel to the sides .
To establish a convenient notation for these figures , he represents the lines in the two directions by numbers separated by a vertical line .
Thus , for example , 310 signifies the mode of vibration , in which there are three lines in one direction and none in the other ; 512 denotes that in which there are five lines parallel to one side , and two to the other , &amp;c. The nodal lines , which may always be considered as having been originally straight , may curve themselves more or less ; and in general the flexions of these lines , whether they adjoin each other or are separated by a straight line , mutually approach to or recede from each other .
In some modes of vibration the nodal lines are never straight .
In some instances , the same mode of vibration may manifest itself in two essentially different ways , according as the flexions of the lines , or the greater number of them , are inward or outward ; in the first case , the sound is usually graver than in the second .
This difference is remarked in those figures where there is an entire number of flexions , as in 210 , 311 , 4110 , 513 , 612 , &amp;c. ; but never in those figures where there are 1 , 2$ , &amp;c. , as in 310 , 411 , 5)0 , 512 , &amp;c. To distinguish the first from the second figure , CHLADNI places a horizontal line above the numbers in the first case , and below them in the second case , thus , 412 , 412 .
When two or more figures having the same notation , occur in the Table , the others are to be considered as distortions of the first , occasioned by altering the fixed points , and the place at which the bow is applied .
If four plates of the same size , and upon which the same figure has been produced , be placed together so as to form a larger square , this compound figure may also be more or less accurately produced on a single larger plate .
Several instances of this may be seen by reference to the Table of figures .
The following Table contains the relative sounds ( expressed both by their musical names and the number of their vibrations , ) of all the modes of vibration of a square plate , experimentally ascertained by CHLADNI .
The horizontal series of numbers denotes the lines parallel to one of the sides , and the vertical series those parallel to the other .
594 + , 343 345 234678 The figures placed as exponents indicate the octaves in which the sounds occur ; and the characters + and denote that the sounds to the characters of which they are affixed are respectively sharper and flatter than the true intervals .
1 595 Having thus briefly stated the general results deduced by CHLADNI from his experimental researches , I shall proceed to class and analyse the phenomena ; and I shall endeavour to show , that all the figures of vibrating surfaces are the resultants of very simple modes of vibration , oscillating isochronously , and superposed upon each other ; the resultant figure varying with the component modes of vibration , the number of the superpositions , and the angles at which they are superposed .
In this first part of the investigation I shall confine myself to the figures of square and other rectangular plates . ?
2 .
The most simple modes of vibration of a rectangular surface are those which exhibit quiescent lines parallel to one of its edges .
EULER has theoretically established , that a rod or band , having both its ends unfixed , can vibrate with 2 , 3 , 4 , 5 , 6 , &amp;c. quiescent lines parallel to the ends , and that the corresponding numbers of vibration are very nearly as the squares of the arithmetical progression 3 , 5 , 9 , 11 , &amp;c. These conclusions are fully confirmed by experimrent .
He has proved , moreover , that when the same mode of vibration of different plates is compared , the number of vibrations is inversely as the square of the length of the plate , but that increase of breadth occasions no difference in the sound ; and that the distance from a free end to a quiescent line is rather less than half the distance between two quiescent lines .
Fig. 1 .
a. page 617 , shows the number and situations of the quiescent lines in the first four modes of vibration of this series .
Fig. 1 .
b. and c. are profiles of the preceding , and represent the curvature of each parallel fibre perpendicular to the quiescent lines at the two opposite limits of their vibration .
The quantity of motion at each point is indicated by the corresponding ordinate of the curve , and its direction by its situation above or below the horizontal line .
It will be convenient to distinguish these states of motion , in which every corresponding point is moving in direct opposition ; and I shall therefore call the first , b. positive states of vibration , and the second , c. negative states of vibration .
When there is an even number of quiescent points , the positive state of vibration may be considered as that in which the motion at the central part is above the plane of equilibrium , and the negative , that in which it is below it .
If we suppose two similar surfaces with the same number of quiescent lines to 596 be superposed , and both to vibrate in concurrence , i. e. both either positively or negatively , they will mutually assist each other 's effects ; but if they vibrate in opposing directions , they will destroy each other 's motions , and the entire surface will be at rest .
3 .
When the rectangular surface is equilateral , it is obvious that it may vibrate in two different rectangular directions , so as to give the same sound , and present the same arrangement of quiescent lines .
Now this plate may be excited at various points where the motion of each motde of vibration is at its maximum , in the same direction , and of equal intensity : such being the case , there is no reason why one mode of vibration should be produced in preference to the other and on calculating the effect of such coexistence , it will be found that the resultants of these combined modes of vibration , similar in everything but in their direction with regard to the sides of the plate , give rise to new quiescent lines which accurately correspond with figures described by CHLADNI ; while the henumber of vibrations does not materially differ from that of the corponent modes of vibration .
The principal results of the superposition of two similar modes of vibration are these : 1st , The points where the quiescent lines of each figure intersect each other , remain quiescent points in the resultant figure ; 2ndly , The quiescent lines of one figure are obliterated when superposed by the vibrating parts of the other ; 3rdly , New quiescent points , which may be called points of compensation , are formed wherever the vibrations in opposite directions neutralize each other ; and , lastly , At all other points the motion is as the sum of the concurring , or the difference of the opposing vibrations .
A primary figure , having an even number of quiescent lines , may be superposed two ways , and may consequently give rise to two distinct resultant figures : one , when the central vibrating parts concur ; and the other , when they are in opposition ; but if the number of the quiescent lines in the primary figure be uneven , there can be only one resultant figure .
The quiescent lines which thus result may be very easily ascertained .
I will take as an example the first mode of vibration , having two parallel quiescent lines ; this being superposed in two rectangular directions , and so that the states of vibration are opposing ( page 617 , fig. 2 .
) , it is obvious that no lines of compensation can exist in the four rectangular segments aaa a , as every point included within them is actuated by concurrent motions ; but in all the other rectangles they must necessarily be formed , as every point within them is affected by the two opposing motions , and if the two modes be of equal intensity , the compensations must occur at every point equally distant from the two rectangular quiescent lines , each appertaining to a different mode of vibration .
The resultant figure will thus be found to consist of two diagonal lines perpendicular to each other , and passing through the centre of the plate .
But if the two superpositions vibrate in concurrence , the rectangles bbbb will be free from compensating points ; but these will occur in the other rectangles , and form the figure represented ( fig. 3 .
) , which also consists of diagonal lines .
In the same manner the resultant of any two similar modes of vibration with nodal lines , parallel to the sides , may be proved to consist of lines parallel to the diagonals . ?
4 .
It is not a necessary condition for the vibrations of a square plate that the primary nodal lines shall be parallel to a side ; they may also be parallel to a diagonal , or to any line intermediate between a transverse and a diagonal line .
In these cases the superpositions take place according to the following rule : That the axes of the superposed modes of vibration must make equal angles with a transversal line passing through the centre ; for otherwise the modes of vibration would not be similar .
By the axis of a primary mode of vibration , I mean a straight line passing through the centre of the plate and parallel to the quiescent lines .
Considerations of the kind already employed will show that in all these instances the resultant figures consist of lines parallel to the edges of the plate , and that they are always the same in number as the nodal lines of a component mode of vibration , but differently distributed in the two directions , according as the angle of superposition varies .
The various primary modes of vibration , transverse , intermediate , and diagonal , and the angles which the quiescent lines of two similar figures make with each other -when they are superposed , are represented in the first column of the general Table , page 619-633 ; in the second column of this Table are placed the figures resulting from their opposing superpositions , and in the third column those which arise from their concurring superpositions .
598 We obtain by experiment a limited number only of figures which can be considered the resultants of primary modes of vibration consisting of any given number of oblique lines ; but it would seem , that as the , various degrees of obliquity are infinite , so there should be an infinite number of resultant figures passing into each other by insensible gradations : by calculation this should be so , but there are causes of limitation which I shall proceed to explain .
It appears that no resultant figure is maintainable unless the greatest excursions of the external vibrating parts occur at the edges of the plate .
In the concurring superpositions of eight oblique lines , this condition can only be fulfilled when the angles they make with each other are either 90 ?
, or 143§ 8 ' ; in the first case the resultant figure consists of four lines in each transverse direction , in the second of six lines in one direction and two in the other .
In the opposing superpositions of the same number of lines , the condition is fulfilled when the angles at which the lines are inclined are 118§ 4 ' and 1630 44 ' ; the resultant figure of the former consists of five lines in one direction and three in the other , and that of the latter of seven in one direction and one in the other , § 5 .
I have , in the preceding sections , described thevarious modes of binary superpositions which may take place on a square surface .
But there are numerous cases in which four superpositions may coexist , and these I shall now proceed to take into consideration .
When the axis of a primary figure corresponds either with a diagonal or with a transverse line , passing through the centre of the plate , it is obvious that there can be only one other line of equal length , which can be considered as the axis of a similar and isochronous mode of vibration ; in these cases it is evident , therefore , that there can only be two superpositions .
But in every intermediate direction of an axis , there are three other lines of equal length which constitute axes of similar modes of vibration ; and four superpositions can therefore take place whenever the axis of a component mode of vibration is neither a diagonal nor a transverse line .
It would be a tedious and laborious process to ascertain a resultant figure by combining its four component modes of vibration ; but the same purpose will be effectually answered by combining them first in pairs , as explained in the 4H 2 599 preceding section , and then combining two of these first resultants rectangularly together .
The following process affords great facility for ascertaining the second resultant figure , which arises from the superposition of two first resultants .
a. and b. ( fig. 4 .
page 617 , ) are the two component first resultants , the similar lines of one being placed rectangularly to those of the other ; the vibrating parts are indicated by the letters P and N , according as the vibrations are positive or negative .
At C the two figures are superposed , A being represented by the continuous and B by the dotted lines .
The surface is now subdivided into a number of unequal rectangles , and by comparing the two component figures together , it is easy to see which of these rectangles are influenced by conspiring , and which by opposing motions ; if the motions are found to conspire , the letters P or N must be placed in these rectangles according as the coexisting motions are positive or negative ; if the motions are in opposition , a mark may be made to indicate that a quiescent line passes through this rectangle .
Wherever a continuous line intersects a dotted line , a mark is to be made , to indicate that a quiescent point is formed ; and as in every other part , the quiescent lines of one figure pass over vibrating parts of the other , the boundary lines of all the rectangles must be marked with the letters indicating the motions of the vibrating parts they superpose .
The figure C being thus marked , the resultant figure is easily described by joining the fixed points by lines drawn through the rectangles shown to be actuated by opposing motions ; carefully avoiding to encroach upon the rectangles of conspiring motion marked P or N. That the diagonal line is perfectly straight may be proved in the following manner .
It must first be premised that the rectangles included within the quiescent lines of each of the first resultant figures , have precisely the same quantity of motion in the same relative points with respect to the surrounding sides , and that the vibrating parts at the edges and corners of the plate must be considered respectively as exact halves and quarters of a complete vibrating part .
This being understood , if two similar first resultants be laid alongside each other in the directions in which they are to be superposed , and if a diagonal line be similarly drawn through each of them , it will be obvious that every corresponding point of each line must possess the same intensity of motion .
If the successive segments ( equal in each figure , ) through which the 600 lines simultaneously pass , be in opposite states of vibration , they will neutralize each other 's effects , and a diagonal quiescent line will be formed ; and if they be concurring , all the parts between the coincident or fixed points will be in motion .
In the diagram ( fig. 4 .
c. ) one diagonal is in the first state and the other in the second .
If the number of lines in the component first resultant figures be uneven , they admit of only one mode of superposition .
But first resultants having an even number of quiescent lines , admit of being superposed in two ways , according as they are vibrating in concurrence or in opposition .
It frequently occurs that entire quiescent lines superpose each other ; thus , when in the component figure there is an uneven number of lines in each direction , the two rectangular central lines of each superposed figure must coalesce , and consequently they continue in both the resultant figures ; exaniples of this are seen in the resultants of 311 , 511 , 513 , 711 , 713 , 715 , &amp;c. Again , when the nunlber of quiescent lines in one direction of the component figure is three times greater than that in the other , the number of coinciding fixed lines in each direction of the resultant figure is equal to the smallest number in the first resultant , as in 612 , 913 , 1214 , 1515 , &amp;c. See the general Table , page 619-633 .
The following general results are obtained from constructing the second resultant figures according to the rules above given .
In superposing first resultants consisting of an even number of lines : 1st , When the number of lines in each direction of the first resultants is even , and the modes of vibration are concurring , no line passes through the centre of the second resultant figure .
When the modes of vibration are opposing , two rectangular diagonal lines of compensation occur .
2ndly , When the number of lines in each direction is uneven , and the modes of vibration are concurring , there are always two perpendicular transversal lines passing through the centre ; and when the modes are opposing , there are , in addition to these fixed lines , the two diagonal lines of compensation .
When the first resultants consist of an uneven number of lines , in which case there is no distinction of concurring and opposing vibrations , one diagonal line only invariably occurs .
In no case is it necessary to calculate an entire figure .
When the number of nodal lines in the primary mode of vibration is even , only one quarter of 601 the figure is required to be calculated , as it is obvious that every second resultant of this kind consists of four symmetrical , and as it were reflected , portions .
But when the primary number of quiescent lines is uneven , it is necessary to calculate one half ; the other half is symmetrical and inverted .
Some of the first resultants are never obtained by experiment .
When the number of quiescent lines in the primary mode of vibration is uneven , either the first or the second resultant may be obtained at pleasure ; thus in 312 if the impulses be made at a corner , where the motion of both superpositions is at its maximum , the second resultant must arise ; but if they be made at the middle of a side , the first resultant only will appear , because the point of excitation is a quiescent point of the other .
But in all cases where there is an even number of lines , it is impossible to obtain the first resultant , because each maximum of vibration equally belongs to both superpositions . ?
6 .
I have given in pages 619-633 , a Table which shows every perfect resultant figure of a square surface when the number of quiescent lines in the primary modes of vibration does not exceed twelve .
I have carefully calculated each figure by the rules laid down in the preceding sections , and have shown in the Table the successive processes of superposition .
The first vertical row exhibits the two primary modes of vibration superposed at the required angles ; one figure being represented by the continuous lines , and the other by the dotted lines .
The second row contains the first resultants which arise from the opposing superpositions of the preceding ; and the third row , those which result from their concurring superpositions .
The fourth and fifth rows exhibit the perfect second resultants which are formed , the former by the opposing and the latter by the concurring superpositions of the first resultant which the plate has been already found competent to produce .
On comparing the calculated figures with those obtained experimentally by CHLADNI , the greater number are found exactly to agree ; there are , however , some differences which it will be necessary to explain .
In the first place , there is an obvious cause of error in delineating figures from experiment , from this circumstance , -that the sand accumulates in the spaces where two convex curves are near and opposite to each other , the motion being there very small , so that it is difficult to ascertain whether the curves join , or not .
Secondly , 602 Inequalities in the plate will sometimes occasion lines which ought to intersect each other , so as to appear separated curves .
On comparing together the figures CHLADNI has marked as 614 , 614 , 712 , 714 , 813 c. &amp;c. with those of the calculated Table , they will be found to differ only in these respects .
Another cause of difference is this : When the lines of one component figure very nearly coincide with those of the other , but without actually doing so , the resultant figure may be such as would arise from their actual superposition , instead of that which accurate calculation would give .
In CHLADNI'S Table there are two instances of this alteration , 712 a. and 813 a. A few of the figures delineated by CHLADNI are irregular resultants formed by the superpositions of dissimilar modes of vibration .
These irregular resultants can be formed only when the dissimilar component modes of vibration give the same sound , and have a maximum point of vibration in common , at which they can be sirnultaneously excited .
The figure marked by CHLADNI 6 1 , I find to be an irregular resultant formed by the combination of 611 with 315 , which both give the same sound C4 .
The irregular figure marked in his Table 511 a. is a compound of 511 and 215 .
The calculated figures 611 , 711 , 811 , 911 , 1011 , 1012 , and 1111 , are not to be found in CHLADNI'S Table .
The near approach of the inclined lines of their primary component figures to parallelism is the cause of the great difficulty in obtaining these figures by experiment .
The figures marked by CHLADNI 1013 , 914 , 815 , 716 , 1014 , 915 , 816 , and 917 , exceed the limits within which I have calculated the Table , and are therefore not to be found in it .
§ 7 .
Imperfect resultant Figures .
I have hitherto considered those resultant figures only which arise from the superpositions of similar modes of vibration , each exactly equal in intensity ; these I shall in future call perfect resultant figures .
But when the vibrations of the superposed modes are unequal in intensity , then a figure intermediate between the perfect resultant and one of its components is formed ; these intermediate figures I shall call imperfect resultants .
They are experimentally obtained by varying in a slight degree the places at which the plate is held or touched , from those necessary to determine the corresponding perfect resultant figure ; the place at which the bow is applied remaining in both cases the same .
Fig. 6 .
a. b. c. d. e. in CHLADNI'S Table , page 611 , represents the successive 603 transformations of figure which take place when each component mode of vibration presents three transversal lines : a. and e. are the two components ; c. the perfect resultant ; b. an imperfect resultant , in which the excursions of a. are the greatest : and d. an imperfect resultant , in which e. has the greatest energy .
Fig. 11 .
a. 6 .
page 611 , exhibits the transitions of the opposing superposition of two primary figures , each presenting four transversal lines ; and fig. 12 .
a. and b. the changes of the concurring superposition of the same .
These are the principal types of the transformation of primary figures into first resultants .
In each of these series of transitions there are certain points which are invariable during every change : these are , the quiescent points formed by the nodal lines of one figure intersecting those of the other , and the centres of vibration where the maxima of positive or negative vibration agree in each component mode of vibration .
The points of compensation are changeable .
Figs. 29 .
a. and 30 .
a. page 613 , represent imperfect second resultants , formed by two superpositions of the first resultant figure 612 .
Fig. 29 .
a. arises from concurring , and fig. 30 .
a. from opposing superpositions .
The straight lines in these imperfect figures arise from the coincidence of entire quiescent lines in each component figure , and , consequently , they remain unaltered whatever may be the relative intensities of the superposed modes of vibration .
But the curved lines , which are formed of compensating points , change with the varying intensities .
8 .
Figures of irregular Plates .
If the sides of the square be nearly , but not exactly , equal , the superpositions of two similar modes of vibration with transversal lines still take place ; but instead of exhibiting perfect resultants , figures resembling transitional figures appear .
Thus in the binary superposition ( page 611 , fig. 1 .
) of the figure with two transversal lines , if the sides be unequal , the crossed lines separate at their point of intersection and are converted into two curves , the summits of which recede from each other as the difference in the lengths of the sides becomes greater .
Also , if the diagonals of the square be unequal , the resultant figure ( page 611 , fig. 2 .
) arising from two superposed modes of vibration with diagonal axes , will be modified in a similar manner .
Corresponding modifications are occasioned , through accidental differences of elasticity , &amp;c. in the directions of the axes of the superposed modes of vibration , even when the dimensions of the plate are apparently equal .
604 If a plate of glass be covered on one of its sides with leaf-gold , or if a plate of ground glass be substituted for an ordinary glass plate with smooth surfaces , the figures may be obtained distinctly delineated by lines consisting of a single row of grains of sand .
Experiments made in this manner , upon square plates carefully prepared , induced Professor STREHLKE to conclude , after many minute measurements of these lines , that all acoustic figures are formed of hyperbolic curves , and that the quiescent lines never intersect each other .
But however correct these experiments may have been , the conclusions drawn from them are unwarranted ; were it possible to obtain plates of a perfectly homogeneous substance , and of accurately equal dimensions , there can be no doubt that the lines , however finely defined , would actually intersect each other .
§ 9 .
I have already , § 1 .
, given CHLADNI'S Table of the comparative sounds , and numbers of vibrations of the figures of square surfaces obtained from experiment .
In the following Table these results are arranged so as to correspond with the views taken in this paper .
The numbers in the first vertical row indicate the number of parallel quiescent lines in the primary figure .
In the horizontal rows the angles at which the lines of the primary modes of vibration intersect each other are shown ; and the figures between brackets give the notation of the first resultant produced by their superposition ; below these its number of vibrations , and the character representing its musical sound , are given ; when there are two of these lower lines separated by a horizontal dash , the one above indicates the sound of the opposing superposition , and that below it the concurring superposition .
Thus it appears , that every figure of a square surface which experiment can give , may be , reduced to a primary figure with parallel lines , giving the same sound : if , therefore , the analytical investigation be confined to these , many of the difficulties will disappear .
EULER has investigated the subject when lines parallel to a side only are concerned ; it remains to extend the inquiry to modes of vibration the nodal lines of which are perpendicular to any line passing through the centre of the surface .
An analytical expression for all the sounds of a square plate may probably be obtained , which shall be a function of the number of quiescent lines , and the length of the axis of the mode of vibration .
Immediately after the publication of CHLADNI'S experiments on square plates , JAMES BERNOUILLI attempted to demonstrate them analytically ; but his investigation was entirely unsuccessful ; his conclusions were founded on erroneous data , and the results he obtained were at variance with experiment .
His assumptions were these : that the primary figures consisted of 2 , 3 , 4 , 5 , 6 , 7 , &amp;c. lines parallel to the side only , these being the modes of vibration of a larina as investigated by EULER ; that any two similar or dissimilar modes of vibration might superpose each other rectangularly , the nodal lines of the two components appearing together in the resultant mode of vibration ; and that the sound of the resultant differed from that of either of the two components , being much higher .
The figures given by experiment he considered accidental distortions of these compound figures .
This theory gave no account of those figures in which a single line in ione direction coexists with any number in the other , as there is o primary figure consisting of one nodal line only ; and BERNOUILLI acknowledged his theory to be imperfect in this respect. .
The failure of BERNOUILLI led CHLADNI inconsiderately to state , that Cthe supposition of regarding such a rigid membraniform body as a network formed by curved lines in one direction applied upon curved lines in another direction , is not conformable to nature , and will never give , either results agreeing with experiment , or an appearance of explanation of some of the most simple vibrations .
" That this assertion is erroneous , the considerations in the present paper have , I conceive , fully proved .
The error of BERNOUILLI did not consist in assuming that the observed acoustic figures were formed by superposing simple modes of vibration on each other , for this has been shown to be true ; but in his assumptions of the manner in which these superpositions were made , and the effects which resulted from such hypothetical superpositions .
The various mathematicians who have more recently undertaken to investigate the laws of vibrating surfaces , as POISSON , CAUCHY , Mademoiselle GERMAIN , &amp;c. , do not appear to have taken into consideration anything resembling the theory of superposition .
Dr. YOUNG seems to have had a correct notion of the origin of the acoustic figures ; for in his Lectures , when slightly noticing CHLADNI'S experiments , he remarks , " The vibrations of plates differ from those of rods in the same man4 2 607 ner as the vibrations of membranes differ from those of chords , the vibrations which cause the plate to bend in different directions being combined with each other , and sometimes occasioning singular modifications .
" The brothers WEBER , in their excellent work the Wellenlehre , published in 1825 , have advanced a step nearer the truth than any of their predecessors .
They have shown , that in a square vessel containg water or mercury , two series of stationary waves , one parallel to each side , may be made to intersect each other ; and that the compound wave formed by their interferences assumes the form shown in § 3 .
to be the resultant of two superpositions of parallel transverse lines .
Their observations are confined to those modes of undulation , analogous to the first resultants of primary modes of vibration with lines parallel to a side .
Though I had advanced considerably in the present inquiry before I saw this work , yet I should be wanting in justice to these philosophers did I not here state that theirs is the merit of having shown , in the most simple case , the way in which the superpositions of modes of vibration or undulation actually do take place .
§ 11 .
Plates of Wood .
Fron the rules already laid down , it is obvious that the series of figures presented by a square plate of any homogeneous material ought not to be obtained on a square plate of wood , in which substance the elasticity is not the same in all directions .
If a square plate of wood be prepared with its fibres parallel to one of the sides of the square , the axes of greatest and least elasticity will be disposed rectangularly , and parallel to the adjacent sides ; in this case the same primary mode of vibration in the two directions will not give the same sound , although the dimensions of the vibrating parts are the same in both ; consequently they cannot coexist , and the resultant figures with diagonal lines will be wanting on such a plate .
But if the axes of the two component modes of vibration be equally inclined to either of the axes of elasticity , these modes of vibration will be necessarily similar and isochronous , and therefore capable of superposition ; on a square plate of wood , consequently , all those first resultants which consist of any number of lines parallel to the sides intersecting each other rectangularly , may be obtained ; and the same figure will be accompanied by different sounds , according as the axes 608 of the modes of vibration are inclined to the axes of least or greatest elasticity .
It is easy to foresee that none of the second resultants , which consist of four isochronous superpositions , can be obtained on such a plate .
But if the wooden plate , instead of being square , be a rectangle , the sides of which are inversely as the squares of their resistance to flexion , the two modes of vibration parallel to the sides , though differing in length , will be isochronous ; and their coexistence will give rise to a resultant figure with lines parallel to the diagonal .
Thus on a rectangular plate of straight-fibred deal wood , in which the proportion of the sides were as 28 to 59 , I obtained the two crossing diagonal lines , corresponding to the second figure of a homogeneous square plate of glass or metal .
SAVART has made a series of numerous and accurate experiments on the changes which take place in the sound , and also in the form and position of the figure of the first mode of vibration on circular plates of wood of similar dimensions , cut in different directions with respect to the three principal axes of elasticity .
All the results he has obtained are in perfect accordance with the rules stated in this paper , and might have been predicted by them .
He has extended his investigations to circular slices of crystals , cut in various directions with respect to their axes , and has obtained in this way much valuable information .
These researches of SAVART point out a new direction to our inquiries respecting the structure of bodies ; and the utility of his experiments will be greatly enhanced by the knowledge we now possess of the causes on which these phenomena depend .
I shall shortly return to this subject .
609 Of all the figures of a square plate , determined by calculation to result from two or four superpositions , when the quiescent lines of the primary figures do not exceed twelve in number .
Primary Figures , showing the angles at which they are superposed to form the first resultants .
90O II .
First Resultants , formed by two superpositions .
Opposing .
Concurring .
111 Second Resultants , formed by four superpositions .
--Opposi .
Concur Opposing .
Concurring .
1 " 0§ 112 III .
180 ?
IV .
I I of Calculated Figures.-(Continued .
) Primary Figures , showing the angles at which they are superposed to form the first resultants .
14 o0 Qt 151 ?
.56 First Resultants , formed by two superpositions .
Opn----- .
Cc-u--rnAOpposing .
Concurring .
Second Resultants , formed by four superpositions .
Oppo .
C Opposing .
Concurring .
114 MDCCCXXXIII .
414 1 , on0 V. 1 Rno VI .
4L of Calculated Figures.-(Continued .
) Primary Figures , showing the angles at which they are superposed to form the first resultants .
1 oo r , o ) First Resultants , formed by two superpositions .
Opposing .
Concurring .
OIA Second Resultants , formed by four superpositions .
Opposing .
Concurring .
157§ 22 ' 180§ 106i§ 16 161§ 41 'L !
4 11 C 4L2 VII .
I of Calculated Figures.-(Continued .
) Primary Figures , showing the angles at which they are surerposed to form the 900 First Resultants , formed by two superpositions .
( " ... . .
T Opposing .
Concurring .
414 Second Resultants , formed by four superpositions .
Opposing .
Concurring .
Opposing .
Concurri [ g. VIII .
118§ 4 ' 117 I lo0 II I : !
of Calculated Figures.-(Continued .
) Primary Figures , showing the angles at which they are superposed to form the first resultants .
1 l(o Alnt 126§ 52 ' IAQO/ ?
CI First Resultants , formed by two superpositions .
Opposing .
C_ .
, _. .
Opposing .
Concurring .
AI .
Second Resultants , formed by four superpositions .
Opposing .
Concurring .
Opposing .
Concurring .
3f6 27 165§ 44 ' IS0§ 9n0 515 IX .
rs1114 -r--r------L L__ _II_ __ ___ __ ____ I , _ ii of Calculated Figures.-(Continued .
) Primary Figures , showing the angles at which they are superposed to form the first resultants .
1120 38 ' 167§ 18 ' First Resultants , formed by two superpositions .
Opposing. .
Concrring Opposing .
Concurring , Second Resultants , formed by four superpositions .
-Oppo .
Concur Opposing .
Concurring .
110 1 rQn 100§ 24 ' XI .
-| , d MDCCCXXXIIII 4MIi - , ~~----~ i--1 7 1r ?
lt _bLiI1 R.0 9OrI 4M of Calculated Figures.-(Continued .
) Primary Figures , showing the angles at which they are superposed to form the first resultant .
1200 30 ' 1380 52t First Resultants , formed by two superpositions .
Opposing .
Concurring .
417 Second Resultants , formed by four superpositions .
Opposing .
Concurring .
318 1680 34 ' 180 ?
XII .
of Calculated Figures.-(Continued .
) Primary Figures , showing the angles at which they are superposed to form the first resultants .
108§ 56t 142§ 9t First Resultants , formed by two superpositions .
Or p , p , C c. i. .
Opposing .
Concurring .
517 Second Resultants , formed by four superpositions .
Opposing .
Concurring .
Opposing .
Concurring .
319 157§ 22 ' 11'11 180 ?
I ... ... ... ?
- ' ... ... I llllll v -fkT~ ' -

