II .
" On the Theory of Probability , applied to Random Straight Lines .
" By M. W. CROFTON , B.A. , of the Royal Military Academy , Woolwich , late Professor of Natural Philosophy in the Queen 's University , Ireland .
Communicated by Prof. SYLVESTER .
Received February 5 , 1868 .
( Abstract .
) This paper relates to the Theory of Local Probability-that is , the application of Probability to geometrical magnitude .
This inquiry seems to have been originated by the great naturalist Buffon , in a celebrated problem proposed and solved by him .
Though the subject has been more than once touched upon by Laplace , yet the remarkable depth and beauty of this new Calculus seem to have been little suspected till within the last 266 [ Feb. 27 , few years , when the attention of several English mathematicians has been directed to it , and results of a most singular character have been obtained .
The problems on Local Probability which have been hitherto treated relate almost exclusively to points taken at random .
The object of the present paper is to show how the Theory of Probability is to be applied to straight lines whose position is unknown , or , in other words , which are taken at random .
The author commences by showing that when a straight line is drawn at random in an indefinite plane , or , in other words , when we take one out of an infinite assemblage of lines all drawn at random in the plane , the true mathematical conception of this assemblage is as follows : Conceive the plane ruled with an infinity of parallels at a constant infinitesimal distance ( 3p ) asunder ; then imagine this system of parallels turned through an infinitesimal angle ( &amp; 0 ) ; then through a second equal angle , and so on , till the parallels return to their original direction ; the plane will thus be covered with an infinite number of systems of parallels , running in every possible direction .
If an infinite plane be covered in this manner with straight lines , and we draw any closed convex contour on the plane , and then imagine all the lines effaced from the plane , except those which meet this contour , we shall have a clear conception of the system of random lines which meet the given contour .
By applying mathematical calculation to this system , the following important principle is proved : The measure of the number of random lines which meet a given closed convex contour is L , the length of the contour .
If the contour be non-convex , or be not closed , the measure will be the length of an endless string passing round it and tightly enveloping it .
Hence , given any closed convex contour of length L , and any other of length 1 , lying wholly within the former , the probability that a line drawn at random to meet L shall also meet 1 , is P=I The following propositions are then established : If the contour I lie wholly outside L , then , if X be the length of an endless band tightly enveloping the two contours and crossing between them , and Y the length of another endless band also enveloping both , but not crossing between them , the probability that a random line meeting L shall also meet 1 , is X-Y -P LL Again , if the contour I should intersect L ( whether in two or more points ) , then , if Y be an endless band tightly enveloping both , L+l-Y P--L 267 A closed convex boundary of any form , of length L , encloses an area 0 : if two random straight lines intersect it , the probability of their intersection lying within it is 27rn The probability of their intersection lying within any given area , w , which is enclosed within Q , is 2ww -L2 .
A more difficult question would be to determine the probability in the case where o is external to &amp; 2 .
These fundamental results , it will be observed , are of great generality .
The author proceeds to apply them to the solution of various problems relating to random straight lines ; in fact any such problem of probability may be reduced by the principles above laid down to a question of pure mathematical calculation .
What will probably be considered among the most curious results contained in this paper are the collateral applications of the theory to the integral calculus .
Several integrals of a singular character are obtained , some of which it seems very difficult to prove by any known method .
One or two of these are subjoined , with indications of the methods used in establishing their truth .
If a given convex boundary be intersected by a system of random lines , as above described , every pair of lines will meet in a point ; and these points of intersection will be scattered all over the plane , some within the boundary , some without .
Those within will evidently be distributed with uniform density over the area ; but it becomes a question for those outside , to determine the law according to which their density varies ; and it is proved in this paper that the density of the intersections of a system of random lines crossing a given area , for any external point P , is proportional to 0-sin 0 , where 0 is the apparent angular magnitude of the area from P. Hence the number of external intersections is represented by ff(0-sin 0 ) dS ; now the number of internal will be wrf , and the whole number LL2 .
Hence If li be any plane area , enclosed by a convex boundary of length L , and 0 the angle which it subtends at any external point ( x , y ) , then f ( Osin 0 ) dxdy= 27r , the integral extending over the whole external sir :face of the plane* .
By conceiving an infinite system of random lines covering an infinite plane , and a second system , all of which meet a given boundary in that plane , and then fixing our attention on the intersections of the former system with the latter , we find the density here proportional to 0 ; and the following theorem is deduced from this consideration : Given any convex boundary ( whose apparent magnitude is called 0 ) , let there be an external boundary surrounding it , such that any tangent to the inner cuts off a constant area from the outer , then ff Odxdy-=D , the integral extending over the whole annulus between them , 1 being the difference of the areas of the parts into which the annulus is divided by any tangent to the inner .
For instance , we may take two similar concentric ellipses .
If both the inner and outer boundaries are of any convex forms whatever , the above expression is still true , provided D mean the average value of the difference of areas as the tangent revolves by uniform angular displacements .
If we consider a plane covered with random lines , and then divide them into two systems , one crossing a given boundary , the other all outside it , the density of the points in which the former system cut the latter will be proportional to sin 0 ; and this leads to the next theorem .
If an endless string ( of length Y ) be passed round a given convex boundary ( of length L ) , and the string be kept stretched by the point of a pencil , which thus traces out an external boundary , then if 0 be the apparent magnitude of the given boundary at any point ( x , y ) , we shall have ffsin Odxdy=L ( Y-L ) , the integral extending over the annular space between the boundaries .
A remarkable instance of this is an ellipse , the outer curve being , as is well known , a confocal ellipse .
Some other applications of the theory to integration are then given .
It is important to notice that these applications , though having arisen from researches on probability , rest on a basis wholly independent of that theory .
The apparatus of equidistant parallels revolving by infinitesimal angular displacements , which has been here employed , is a purely geometrical conception ; and the proofs of these integrals can be presented in a strict mathematical form .
A reticulation composed of two systems of parallels crossing at a finite angle has already been employed by Cauchy , Liouville , and Eisenstein as a method in the theory of numbers and elliptic functions .
The reticulation used above is a more delicate and complicated one , consisting , not of two , but of an infinite number of systems of parallels .
There remains a more difficult but deeply interesting inquiry , scarcely touched upon in this paper-namely , the extension of the above results to the cases of straight lines , and of planes , taken at random in space .

