]\gt ; Mr. W. K. Edwards .
" " A Radial rea-Scale.\ldquo ; By W. K. EDWABDS , M.A. municated by Professor A. G. GREENHILL , F.R.S. January 28 , \mdash ; Read March 3 , 1904 .
This is a contrivance for finding the area of a plane figure by of a transparency .
The design in the transparency consists number of radiating lines .
Each of these lines is graduated .
There are various patterns of this design , and their respective to convenience and accuracy form a wids field for disoussion .
In accompanying transparency reproduced in the figure ) , which is simple and effective , there are eleven straight lines radiating from point at equal angles .
The way in which the transparenoy is used as follows :\mdash ; The figure whose area is to be found is placed under the close contact with it , so that its contour lies just between the outside lines of the transparency , i.e. , so that each outside line the contour , or passes through a cusp or angular point , or some rectilineal portion of the contour .
Each of the radiating thus becomes a or transversal , or contains a side of the The graduations of the right-hand points of intersection of transversaIs are read and added together ; then the graduations of left-hand points of intersection are read and added together .
second sum is subtracted from the first ; and the resuIt records number of square inches in the fig tlre .
It will be seen that if each of the outside lines touches , or through an angular point of the figure , there will be eighteen tions to be read\mdash ; those on the outside lines cancelling each other .
one of the outside lines contains a rectilineal portion of the there will be twenty graduations to read ; if both ontside lines do twenty-two graduations be recorded .
If a quicker use of the area-scale , with less ance of accuracy , is desired , the figure to be quadrated may be placed so as to lie just between the first and ninth , or the first and seventh , or the first and fifth lines , in which cases fourteen , or ten , or six graduations will be read respectively .
If the figure is too large to be included between the outside lines , it may be divided two parts by a straight line drawn across it , or into three parts by a pair of lines inclined to one another at the same angle as the outside lines , which may be done by means of the cardboard slip accompanying the diagram .
The second of these methods of dividing up the figure may be also employed when it is desired to obviate the inaccuracy that may result from the two outside lines otherwise being tangent to the figure .
Mr. W. K. Edwards .
Principle of the Suppose the equation of a curve to be the nrea ) etween the curve , the initial line , and the radius anglc with the initial line , is ; i.e. , , say .
Now if be the outside radii of the sector so quadrated , and the vcctor isecting the angle between them , it can be seen ; whence , assuming , and solving three of four simultaneous equations for , we get vahles A which also satisfy the fourth .
We have then The design on this particular is made on the tion that each of the tell separate portions of the curve between first and third , the third and fifth , . . .
the ninth and radiating lines , approximates to some member of the above family curves .
Anyolle acquainted the of , in Cartesit ) ordinates , in terms of a series of equidistant ordinates , common distance apart , given in Bertrand 's Calcul Integral , Sectionl will see analogy in the foregoing expression of the of sector in terms of the squares of three equally inolined radii neir common inclination .
Other designs of the transparency can made on the assumption of larger portions of the cnrve to members of families of in which is equal to a algebraical function of in ascending powers of , of degree than the third .
On the whole it is easier , and possibly more to use the method employed above , in the same way as Rule " " uses the corresponding theorem Cartesian There is no reason to suppose that " " Simpson 's Rule\ldquo ; gives a accurate approximation in the generality of oases than Rule\ldquo ; or the numerous rules deducible in ternls of the calculated by Cotes .
An investigation of the family of curves * Sec Boole 's ' liinite Differences .
' A that they form a largo variety of spirals .
An important point notice is that if , as usual , dellotes the inclination of a radius to tangent at its extremity , we ) , , equal to in ge1leral , cannot ) zero lmless at the , when the curve th ough the origin .
This means that error is likely to be when the curve is such that any tangent to it through the point of radiation ; , when the outside lines are to the .
or the nrve is re-entrant in a } any one of the radiating lines gives three or four readings .
The may be expected to give the best resuIts , therefore , when so that the outside past throngh cnsps or angular points the curve , and so that ting line crosses it more th curves in which there are no pointq , it is best , therefore , to into two or three areas .
There is no reason why the presence of of flexure be supposed to vitiate the resuIts .
For the equation for the positions at which such points occur in the stated of will be found to be , in general , of the sixth degree .
deviations from strict may be expected in the of nearly all the -known curves by means the area-scale .
If , as in the present , the angle between the lines is half a radian , these tions will be found to be except in the of oval the outside and not treated as suggested 011 p. 292 .

