XXIII .
The general Mathematical Laws which regulate and extend Proportion universally ; or , a Method of comparing Magnitudes of any Kind together , in all the possible Degrees of Increase and Decrease .
By James Glenie , A. M. and Lieutenant in the Royal Regiment of Artillery .
THE doctrine of proportion laid down by EUCLID , and the application of it given by him in his Elements , form the basis of almost all the geometrical reasoning made use of by mathematicians both ancient and modern .
But the reasonings of geometers with regard to proportional magnitudes have seldom been carried beyond the triplicate ratio , which is the proportion that similar solids have to one another when referred to their homologous linear dimensions .
This boundary , however , comprehends but a very limited portion of universal comparison , and almost vanishes into nothing when referred to that endless variety of relations , which must necessarily take place between geometrical magnitudes , in the infinite possible degrees of increase and decrease .
The first of these takes in but a very Read March 6 , a very contracted field of geometrical comparison ; whereas the last extends it indefinitely .
Within the narrow compass of the first , the ancient geometers performed wonders , and their labours have been pushed still farther by the ingenuity and indefatigable industry of the moderns .
But no author , that I have been able to meet with , gives the least hint or information with regard to any general method of expressing geometrically , when any two magnitudes of the same kind are given , what degree of augmentation or diminution any one of these magnitudes must undergo , in order to have to the other any multiplicate or sub-multiplicate ratio of these magnitudes in their given state ; or any such ratio of them as is denoted by fractions or surds ; or ( to speak still more generally ) a ratio which has , to the ratio of the first-mentioned of these magnitudes to the other , the ratio of any two magnitudes whatever of the same but of any kind .
Neither have I been able to find that any author has shown geometrically in a general way , when any number of ratios are to be compounded or decompounded with a given ratio , how much either of the magnitudes in the given ratio is to be augmented or diminished , in order to have to the other a ratio , which is equal to the given ratio , compounded or decompounded with the other ratios .
To investigate all these geomeN nn2 trically , trically , and to fix general laws in relation to them , is the object of this paper ; which , as it treats of a subject as new as it is general , I flatter myself , will not prove unacceptable to this learned Society .
It would be altogether superfluous for me to mention the great advantages that must necessarily accrue to mathematics in general , from an accurate investigation of this subject , since its influence extends more or less to every branch of abstract science , when any data can be ascertained for reasoning from .
I shall , in a subsequent paper , take an opportunity of showing how , from the theorems afterwards delivered in this , a method of reasoning with finite magnitudes , geometrically , may be derived , without any consideration of motion or velocity , applicable to every thing to which fluxions have been applied ; and shall now proceed to the subject of this paper , after premising the two following definitions .
DEFINITION I. Magnitude is that which admits of increase or decrease .
DEFINITION II .
Quantity is the degree of Magnitude .
By magnitude , besides extension , I mean every thing which admits of more or less , or what can be increased or diminished , such as ratios , velocities , powers , &amp;c. As I shall frequently , for the sake of conciseness and convenience , be obliged to make use of particular modes of expressing geometrical magnitudes , I here observe , once for all , that by such expressions as these A. A/ B , A. A-B/ B , A. C-D/ D , A.A-B~2/ B , &amp;c. I mean respectively a third proportional to B and A ; a fourth proportional to B , A , and the difference of A and B ; a fourth proportional to D , A , and the difference of C and D ; a fourth proportional to B , A. A-B/ B and A-B , &amp;c. To proceed then in the order in which I first investigated these theorems ; let A , B , C , D , E , F , G , H , &amp;c. be any number of magnitudes of the same kind , taken two YZSR ~IXTQP ECFD A. C-D/ D AB VKMLMNO and two from the first ; and let MN , NO , NR , OP , respectively represent A , B , C , D. Let NR , OP , be drawn perpendicularly pendicularly to VO , or otherwise if in the same angle ; and let the rectangles or parallelograms MR , NP , be completed .
Let LM be a fourth proportional to OP , MN and NR OP ; and let the rectangle or parallelogram LQ be completed .
Then ( 14 .
E. 6 .
) LT is equal to TR , and consequently LQ to MR. But ( 23 .
E. 6 .
) MR has to NP the ratio compounded of the ratios of MN to NO and NR to OP .
Therefore ( 1 .
E. 6 .
) LN has to NO the ratio compounded of the ratios of MN to NO and NR to OP .
But LN is equal to MN + MN .
NR-OP/ OP , or A+ A. C-D/ D , by construction .
Whence it appears , that a magnitude of the same kind with A and B , which has to B the ratio compounded of the ratios of A to B and C to D , is expressed by A+ A. C-D/ D. In like manner let E , F , be represented by RN , OP , respectively , and let LK be a fourth proportional to OP , LN , and QR .
Then ( 14 .
E. 6 .
) KX is equal to XR or TR and XS together .
But since LN hath already been shown to be equal to A+ A. C-D/ D , LK is a fourth proportional to F , E-F , and A+ A. C-D/ D ; that is equal to A. E-F/ F + A. C-D/ D. E-F/ F. by construction .
Wherefore KN being equal to LK + LN is equal to A+ A. C-D/ D + A. E-F/ F + A. C-D/ D. E-F/ F. And since KQ is equal to LR , KN has to NO a ratio compounded of the ratios of LN to NO and NR to OP ; that is , of the ratios A to B , C to D , and E to F. Therefore a magnitude of the same kind with A and B , which has to B the ratio compounded of these ratios is expressed by A+ A. C-D/ D + A. E-F/ F + A. C-D/ D. E-F/ F. Again , if NR , OP , be supposed to represent G , H , respectively , and KV a fourth proportional to OP , KN , and QR ; VQ will be equal to KR ( 14 .
E. 6 .
) and consequently UN will have to NO a ratio compounded of the ratios of KN to NO and NR to OP ; that is , of the ratios A to B , C to D , E to F , G to H. But VK is by construction equal to A. G-H/ H + A. C-D/ D. G-H/ H + A. E-F/ F. G-H/ H + A. C-D/ D. E-F/ F. G-H/ H. And this added to KN above found gives A+ A. C-D/ D + A. E-F/ F + A. G-H/ H + A. C-D/ D. E-F/ F + A. C-D/ D. G-H/ H + A. E-F/ F. G-H/ H + A. C-D/ D. E-F/ F. G-H/ H , for the magnitude of the same kind with A and B , which has to B the ratio compounded of the ratios A to B , C to D , E to F , G to H ; whence the law of continuation is manifest .
The same conclusions may be derived from ( E. 5 .
) ; so that no principle can be simpler or more geometrical than that here made use of .
Thus then these magnitudes will stand .
1 .
A+ A. C-D/ D , when two ratios are compounded .
2 .
A+ A. C-D/ D + A. E-F/ F + A. C-D/ D. E-F/ F , when three are compounded .
3 .
A+ A. C-D/ D + A. E-F/ F + A. G-H/ H + A. C-D/ D. G-F/ F + A. C-D./ D G-H/ H + A. E-F/ F. G-H/ H + A. C-D/ D. E-F/ F. G-H/ H , when four ratios are compounded , &amp;c. &amp;c. By continuing this operation much farther , I found upon examination that the number of terms in which A is connected with the differences C-D , E-F , G-H , &amp;c. taken one by one , two by two , three by three , &amp;c. if p denote the number of ratios compounded , is expressed respectively by p-1/ 1 , p-1/ 1 .
p-2/ 2 , p-1/ 1 .
p-2/ 2 , p-3/ 3 , &amp;c. Thus if the ratio of A to B be supposed equal to the ratios of C to D , E to F , G to H , &amp;c. respectively , these expressions will give the following ones .
1 .
A+ 2-1/ 1 .
A. A-B/ B. 2 .
A+ 3-1/ 1 .
A. A-B/ B + 3-1/ 1 .
3-2/ 2 .
A. A-B~2/ B. 3 .
A+ 4-1/ 1 .
A. A-B/ B + 4-1/ 1 .
4-2/ 2 .
A. A-B~2/ B + 4-1/ 1 .
4-2/ 2 .
4-3/ 3 A. A-B~3/ B ; for magnitudes of the same kind with A and B , which have to B respectively the duplicate , triplicate , and quadruplicate ratio of A to B ; where p is successively equal to 2 , 3 , and 4 .
And universally , by the same geometrical reasoning , it is found , that A+ p-1/ 1 .
A. A-B/ B + p-1/ 1 .
p-2/ 2 .
A. A-B~2/ B + &amp;c. A. A-B~p-1/ B has to B such a multiplicate ratio of A to B as is expressed by the number p. In the reasoning above I fixed on B as the magnitude to which the rest were to be referred ; but I might as well have fixed on A or any of the other magnitudes .
Thus , for instance , B+ p-1/ 1 .
B. B-A/ A + p-1/ 1 .
p-2/ 2 .
B. B-A~2/ A + &amp;c. B. B-A~p-1/ A has to A such a multiplicate ratio of B to A as is expressed by the number p ; or A has to B+ p-1/ 1 .
B. B-A/ A + p-1/ 1 .
p-2/ 2 .
B. B-A~2/ A + &amp;c. B. B-A~p-1/ A the ratio of A+ p-1/ 1 .
A. A-B/ B + p-1/ 1 .
p-2/ 2 .
A. A-B~2/ B + &amp;c. A. A-B~p-1/ B to B ; that is , such a multiplicate ratio of A to B as is expressed by the number p. Each of these , indeed , I demonstrated separately from the same sort of geometrical reasoning ; but for the sake of brevity I omit setting down these separate demonstrations , as they are both contained in general reasoning above , which furnishes likewise a great variety of other expressions , according as certain numbers of the ratios C to D , E to F , G to H , &amp;c. are supposed to be respectively equal to , greater or less than , the ratio of A to B.

