]\gt ; 1879 .
] Spontaneo us Segme tation of a Liquid By drawing lines through all the positions of each particular metal in the tables of solutions of different strengtbs of either of the saIts employed , it was found that the position of pairs of metals of like properties , such as magnesium and zinc , nickel and cobalt , rhodium and iridium or platinum , lead and tin , \amp ; c. , often varied , and described similar curves or lines of variation .
A much greater number of variations of relative electric position of the metals occurred in the cold and hot solutions of potassic cyanide than in either those of the chloride , bromide , or iodide , whether cold or hot , and the amounts of such variations were also much greater in the cyanide solutions .
Agitation of the immersed metals is well known to influence the current .
Shaking one of the metals reversed occasionalIy the direction of the current ; palladinm was the metal which most frequcnfly manifested this effect .
The agitation of that metal rendered it either more negative or less positive in a considerable degree , especially in solutions of potassic chloride and cyanide .
In those of the bromide , antimony was the metal { he current from which was most aifocted by shaking .
The effect of shaking was less in the iodide solutions tball in those of the bromide .
VIII .
' On the Spontaneous Segmentation of a Liquid lus By A. M. WORTHINGTON , M.A. Communicated BALFOUR STEWART , F.R.S. , Professor of Natural Philosophy in Owens College , Manchester .
Received December 6 , 1879 .
In seeking an explanation of the appearance of a definite number of lobes in the liquid annulus which is formed at a certain stage of the splash of a drop that has fallen vertically on to a horizontal plate , I was led to make some experiments on the ) ntaneous segmentation of such an annulus lying 011 a plate with a view to ascertaining whether the relation that exists between the d'inensions of the annulus and the number of drops into whiclr it will spontaneously split .
is the same as for a straight cylinder of liquid under similar conditions .
Finding that the relation was the sa11le , it then occurred to me that by liberating an annulus in air without contact with any solid , direct experimental proof mightbe obtained of the law of segmentation of a free cylinder .
It has been shown mathematically in various ways , and experimentally by M. Plateau , that the equilibrium of a free cylinder of any liquid , under the influence of surface tension only , becomes unstablo as soon as the length exceeds times the diameter ; and it has beolt legarded as a necessary consequence of this that such a cylindor , if VOL. XXX .
Mr. A. M. Worthington on the once realised , will spontaneolusly split into as many equal and equidistant spheres as times the diameter is contained in ths length .
Thus , if be the number of drops , the length , and the diameter or thickness of the cylinder , the law of segmentation is expressed by the formula M. Plateau has endeavoured to ] an experimental proof of this by liberating a cylinder of mercury , extended between two amaIgamated wires .
on a glass plate , but he found the frictional contact with the plane surface along the length of the cylinder diminished the number of drops , probably because the limits of stability are extended by friction : thus , when the cylinder of mercury rested on a glass plate , the limit of stability appeared to be extended to times the diameter , being the average distance apart at which the drops were formed .
So that the law became , and it was only by observing the effect of simuItaneous contact with a vertical as well as a horizontal plate that a rough estimate could be made of the lence of friction on the number of drops .
By the method of liberation in air that is described in the latter half of this paper all friction is avoided , and the results show a close accordance with theory .
I will first describe experiments with reference to an annulus lying on a plate .
It may be well at the outset to remark that an a , nnulus is not , like a cylinder , a figure of equilibrium for a liquid , there being excess of pressure on the outside ; this excess , however , will diminish as the radius of the annulus increases , \mdash ; as , in fact , it approximates to a cylinder .
The excess , moreover , will not be a large fraction of the total surface.pressure , if , as in the case have to deal with , the thickness of the annulus is small in comparison with its width : thus in the figure\mdash ; the principAl radii of curvature at A are and and at are and .
1879 .
] pontaneous Segmentation of a Liquid Annulus . .
pressure at and at where is a constant , and the difference of pressures at A is proportional to , which is small compared with It may , therefore , well happen , as we shall see in the sequel that it does , that the annulus will split spontaneously before any appreciable alteration of its form can be caused by this inequality of pressures ; and tlJis is the more to be expected if the annulus be lying on a horizontal plate , friction with which impedes the motion of the liquid as a mass .
The method that I have adopted to discover the law of splitting has been to liberate suddenly an annulus of mercury on a horizontal surface of glass or iron .
The apparatus consisted of a circular plate of iron 15 millims. thick and 60 millims. in diameter , the lower surface of which was truly plane , and in it were cut various annular grooves , whose cross section was an exact semicircle .
This plate is placed face downwards on a plane horizontal surface of iron or glass , and mercury is poured into one of the grooves through one of two small vertical holes bored through the plate , the second hole allowing the imprisoned air to escape .
After removing any excess of mercury that may remain in the holes , the plate is suddenly lifted by a central handle , and a circle of regularly disposed drops is found on the lower plate .
The following is a table of resuIts obtained in this way , which are recorded without any omissions .
The size of the grooves was obtain by direct measurement and by weighing the mercury they contained .
A groove of 7 millim. thickness was found to be too small to use , being very difficult to fill with mercury .
1879 .
] Spontaneous Segmentation of a Liquid Annulus .
3rd Groove .
Diameter millims. Principal radius In this case the quantity of liquid was so great that it was found very difficult to lift the plate without giving it an initial disturbauce , which caused some of the drops to run together ; hence it was impossible to lift the plate as quick .
] as had been found necessary in th previous cases to secure the maximum mlmber of drops* Before discussing these resuIts , I may remark that the number of drops was practically the same , whether on iron or glass , unless , deed , the iron was very well rubbed with a.dry clean cloth betweell each experimenb .
When this was done the number of drops was slightly increased .
On one occasion 21 were obtained from the first groove .
In many instances subordinate secondary drops , due to the splitting of the long necks joining the drops before their complete separation , were formed with great regularity ; bnt sometimes these secondary drops were so large as to be easily mistakeu for primary drops .
the case of the third groove , it required practice to lift the plate quickly enough to secure the maximum number of drops .
Mr. A. M. Worthington on the [ Dec. 18 , In what follows I shall , for convenience , call the figure that would be formed by a cylinder into an annulus , a cylindrical annulus , while the form in which the liquid that has filled a groOYe is left on the lower plate may be cafled a semi-cylindrical annulus .
It seems reasonable to suppose that the first thing that will happen to the liquid when liberated will be that it will spring from a semicylindrical into a more or less cylindrical annulus , for there is a sharp curvature of the surface at the edges bounding the flat and curved portions , which will at once exert a pressure on the liquid , while the force which causes it to split into drops being due to an unstable equilibrium*must begin with a value equal to zero , and will therefore be behindbanCL in its action .
Since the cylindrical annulus must have the same volume as the cylindrical , its diameter must be diameter of the semiannulus according to previous notation .
Hence , if we apply Plateau 's formula , for the cylinder , we have\mdash ; 6 .
which gives for the first groove , drops .
second , , third , , The only notable deviation from the number actually observed , is in the case of the third groove , but even here is within the limits of ervation , and a very possible value , since the plate , as has been explained , could not be lifted up quickly enough to avoid all friction .
The observations , therefore , prove that ' at least for an annulus of mercury lying on a ] plate , the law of splitting is practically the same as that of a straight cylinder of the same thickness , being expressed by the formula PART II .
In order to effect the liberation of a free amnulus in air , the apparatus shown in Iig .
1 was used .
A is a circular disk of hard wood , in which two annular grooves were cut , whose cross section was of the form shown in fig. 2 , the lower part being The width of the aner groove at the symmetry of the annulus implies equihbrium as regards segmentation .
Spontaneous Segmentation of Liquid .
55 top was 4 centim. , and of the outer 625 centim. ; the principal diameters being 5.35 centims. and centims. respectively .
is a sort of catapnlt to project the disk vertically downwards , and consists a short tube 3 cenbims .
long , clamped vertically , in which an as inner tube slides locsely , but is kept thrust out as far as the slots the outer tube permit , by the action of a strong spiral spring inside .
By looping a string , , over a small central hook in the disk , the latter can be drawn up tight against the end of the outer tube , held in a horizontal position , and then , after a known weight of mercury has been poured into either groove , the end of the string is let go , and the disk is shot downwards , leaving the mercury annulus in the air .
A tray of fine wet sand is placed below the annulus to catch and preserve the impressions of the circle of drops formed , while the disk is steered out of the way by striking with its fwo projecting arms against the ' brass rods , as showtl in the figure .
The inclination of these rods should be such as to get the disk out of the way as soon as possible , and will depend on the strength of tho spring used ; the stronger the better .
The object of slanting the sides of the grooves is to prevent the annulus being carried downwards by amy friction with the disk .
1 have tried smoked glass , lard , and wet sand , for receiving the drops and their impressions \mdash ; the last is cleanest and answers best ; the best consistence is easily found by practice , and depends a little on the depth of fall .
If too soft , the impressions of separate drops 56 Mr. A. M. Worthington on the mingle , too hard , drops are liable to rebound from their shallow impressions , and to mingle with the neighbouring drops .
If the tray of sand be placed at any considerable distance below the starting point of the annulus , it will be that the impression made is a single small irregular pit with three or four lobe-like projections , showing that the tnulus has contracted as it fell into a single not perfectly symmetrical mass .
This is the natural consequence of excess of on the outside circumference , which is no longer hindered by any as in the eceding experiments ; it is interesting to observe , as the depth of fall is diminished , the diamBter of the .
impression lncreases , while the number of separate Mop-impressions likewise lncreases , showing that the contraction continues , by impeIns , after complete segmentation has taken place , .
amd causes the re-fusion of drops that have at one time been separate ; this re-fusi\copy ; n being gene ] ally so conducted that the symmetrical arrangement of the is tolerably well This will be seen an inspection of tIoe following tables of observations , which are recorded without omissions .
A few remarks on the tables are necessary- It often happened that ] had not remained circular , but had assum a more or less elliptical form , as shown by the impressions ; the tables contain the measures of the greatest and least diameters in such cases , as well as their mean .
These measurements were made with a pair of compasses , and the diameters mentioned joined what were to be tho middle points of the drop marks on either side of the centre ; the measures may be taken as correct millim. , but not beyond .
In most cases the impressions were not equally distinct romnd the whole of the circumference , but were most distinct and well divided itbout one end of a diameter , and least so about the other end of the same .
When the mark was elliptical thia diameter was generally the shortest .
Owing to the fact tbalt each drop did not remain in the hole it bad made in tfJe s and that the partitionabetween neighbouring drop-impressions .
were often cut by the neck of liqmd that remains till mentation is complete , it could not be positively asserted in all cases ( though it could in some ) that the annultls had splib into such and such number of drops when it reached the sand at any given depth , but rather that it consisted of a certain countable number of masses of merc , , some of which appeared to consist of two , or in some cases , of three drops , still united by necks of liquid , while the remaining majority consisted of a countable nIlmber of separate apparently equal and regularly disposed drops .
Then there in many cases a ht uncertainty as to the exact number , to the extent of one drop more or less , and this could only be It will be observed that , in the case of the smaller annulus , the experimental value of is nearly a whole drop below the tbeoretical value .
This points to the possibility of contraction before the segmentations were determined .
Snch contraction would cause a diminution in the value of the ratio , bobh by diminishing the numerator and incl.easing the denominator of the fraction , and would produce a corresponding diminution in the number of drops .
Since the force of contraction is inversely proportional to the principal radius of the , its effect will be more rapid in the smaller annulus , both on this account and because of the smaller mass of liquid to be set in motion .
* That the annulus sometimes becomes elliptical , is probably due to an uneven distribution of the liquid at starting .
When the disk was llot quite horizontal the thickest and thinnest portions of the annulus would be opposite to each other , and the segmentation of the latter will be complete before that of the former .
This accords with the previous observation that the well and ill-divided parts are always at opposite ends of the same diameter .
To account for the fact that this diameter is the shortest , it would sbould be remembered that the contractive force ceases to act on drops which have assumed the spherical form .
Hence the total amount of contraction that is found to have taken place at time after complete segmentation depend ou the velocity acquired the force acted , and therefore upon the time required for the process of segmentation .
This time will evidently be less in the of a thinner annulus .
60 Spontaneous tation of a Liquid Annulus .
[ Dec. 18 , be necessary to evaluate the average centripetal velocities during the fall of the portions of the annulus at the positions of maximum , mean , and mlnlmum thickness , for a given distribution of the liquid , which can only be done when the connexion between time of segmentation and thickness is known .
The very fact of segmentation taking place sooner in one portion of the annulus than anokher is itseIf a source of asymmetry , as was shown by the experiment of liberating a portion of an annulus , which was found to contract in length and straighten itseIf with great rapidity .
Thus after a fall of 12 , a portion of annulus No. 1 , represented in flg .
8 , gave on the sand the impression )resented in fig. 4 .
The sharp curvature of the free ends is of course the cause of the shortenin .
Its influence was often noticed in the case of elliptical marks , masmuch as the extremities.of the row of ill-divided impressions were separated from the remi ties of the well-divided portion by spaces rather greater distance between neighbouring drops eIsewhere .
It is evident that the method of beration adopted in these experiments is not applicable to a straight cylinder , which will immediately begin to contract and under the influence of pressure of the curved ends , unless it be prevented from doing so by being kept stretched between metallic terminals to which it adheres ; terminals would , unless of reat mass , require to be fixed to a frame which of course fall with cylinder .
The impurity introduced by using an amalgamated metal would be a source of error , as M. Plateau has pointed out in his own experiments , which however , might be avoided by adopting the plan mentioned by Professor Guthrie , " " Proc. \ldquo ; vol. xiii , p. 458 , of platinum , to which , when once rubbed with an amalgam of sodium , mercury will dhere closely and without amalgamation .
The experiment would , in this manner , be probably quite feasible .
I have endeavoured by smoking the wooden disk , to render it unwetable by water , and so liberate an annulus of that liquid , but so far without success ; the water always finds its way here and there through the coating of lamp-black and adheres to the disk .

