TRANSACTIONS .
I. Calculations to determine at what Point in the Side of a Hill its Attraction will be the greatest , &amp;c. By Charles Hutton , LL. D. and F. R. S. In a Letter to Nevil Maskelyne , D. D. F. R. S. and Astronomer Royal .
Read Nov. 11 , 1779 .
DEAR SIR , Royal Military Acad. Sept. 21 , 1779 .
AS the experiment of determining the universal attraction of matter , which you lately conducted with so much accuracy and success , is of so great imVOL .
LXX .
B portance portance and curiosity , that it may probably be hereafter repeated by the learned in other countries , and in other situations ; and as the utmost precision is desirable in so nice an experiment ; I have composed the enclosed paper to determine the best part of a hill for making the observations , so as to obtain the greatest quantity of attraction .
I have no doubt , Sir , that the determination of this point will appear of some consequence in the opinion of one who has the improvement of useful knowledge so much at heart ; and if the manner in which it is here made , meet your approbation , I shall desire the favour of your presenting it to the Royal Society .
I have the honour to be , &amp;c. 1 .
The great success of the experiment , lately made by the Royal Society , on the hill Schehallien , to determine the universal attraction of matter , and the important consequences that have resulted from it , may probably give occasion to other experiments of the same kind to be made elsewhere : and as all possible means of accuracy and facility are to be desired in so delicate and laborious an undertaking ; it has occurred to me that it might not be unuseful to add , by way of supplement to my paper of calculations relative to the above-mentioned experiment , an investigation of the height above the bottom of a hill , at which its horizontal attraction shall be the greatest ; since that is the height at which commonly the observations ought to be made .
and since this best point of observation has never been any where determined that I know of , but has been variously spoken of or guessed at , it being sometimes accounted at 1/ 3 , and sometimes at 1/ 2 of the height of the hill ; whereas from these investigations it is found to be generally at about only 1/ 4 of the altitude from the bottom .
2 .
Let ABCEDA be part of a cuneus of matter , its sides or faces being the two similar right-angled triangles ABC , ADE meeting in the point A , and forming the indefinitely small angle BAD .
Then of any section beed , perpendicular to the planes ABD E C and ADE , the attraction on a body at e c A in the direction AB , is equal to the dD constant quantity Ss ; where S= sin. AbB &lt; BAC and s= sin. &lt; BAD , to the radius 1 .
For , first , since the magnitude of the flowing section is every where as Ab2 , and the attraction of the particles of matter inversely as the same , or as 1/ Ab2 ; therefore their product or Ab2/ Ab2 or 1 ( a constant quantity ) is as the force of attraction of beed .
Then to find what that quantity is .
Put AB = a , and BC = x ; then BD or CE ( the distance between the two planes at the distance AB ) is = as .
Now the force of a particle in the line CE is as 1/ AC2 in the direction AC , and therefore it is as AB/ AC3 in the direction AB ; consequently the force of the whole lineola CE in the direction AB is AB .
CE/ AC3 ; and therefore the fluxion of the force of the section BCED or f is = AB .
CE/ AC3 .
BC = a.as.x/ ~a2 + x~ = a2sx/ ~a2 + x2 3/ 2 ; and the fluent gives f= sx/ ~a2 + x2 =sX BC/ AC = sS for the attraction itself .
3 .
To find now the attraction of the whole rightangled cuneus on a body at A in the direction AB .
Since the force of each section is Ss by the last article , therefore the force of all the sections , the number of them being AB or a , is asS = s. AB .
BC/ AC the force of the whole cuneus ABCEDA .
4 .
To find the attraction of the rectA B angular part ABCD on A in the direction AB ; ABCD being one side of the cuneD C us , and AD its edge.-Put AD = BC = b , and AB = x. Then , the force of any section as BC being always as Ss by Art .
2 , the fluxion of the force or f will be = Ssx = sx X BC/ AC = sx X b/ ~b2 + x2 = bsx/ ~b2 + x2 ; and the fluent is f= bs X hyp .
log .
of x+ ~b2 + x2/ b = s. BC X hyp .
log .
AB + AC/ BC = the attraction of ABCD .
5 .
To find the attraction of the C right-angled part BCD of a cuneus Q whose edge passes through A the ADPB place of the body attracted.-Put AB = a , BC = b , BD = c , DA =d=a c , DC = e , AC = g , and DP = x. Then , the force of any section PQ being still as Ss , the fluxion of the force of the part DPQ is f = Ssx = sx X PQ/ AQ = bsxx/ ~c2d2 + 2c2dx + e2x2 ; and the correct fluent when x=c is f= bcs/ ee Xgd dc/ e X hyp .
log .
ee + eg + dc/ de + dc = the force of a body at A in the direction AB .
6 .
To find the attraction of the C right-angled part BCD on the point Q A. Using here again the same ABPD notation as in the last article , we have f= Ssx = sx X PQ/ AQ = bsxx/ ~c2d2 2c2dx + c2x2 .
The correct fluent of which , when x= c , is f= bcs/ ee Xgd+ dc/ e X hyp .
log .
ce + eg dc/ de dc .
7 .
To apply now these premises to the finding of the place where the attraction of a hill is greatest , it will be necessary to suppose the hill to have some certain figure .
That position is most convenient for observing the attraction , in which the hill is most extended in the east and west direction .
Supposing then such a position of a hill , and that it is also of a uniform height and meridional section throughout ; the point of observation must evidently be equally distant from the two ends .
But instead of being only considerably extended , I shall suppose the hill to be indefinitely extended to the east and to the west of the point of observation , in order that the investigation may be mathematically true , and yet at the same time sufficiently exact for the beforesaid limited extent also .
It will also come nearest to the practical experiment , to suppose the hill to be a long triangular prism , so that all its meridional sections may be similar triangles .
Let therefore the triangle ABC represent its section by a vertiB cal plane passing through the meridian , or one side of an HPDEF indefinitely thin cuneus whose edge is in PG ; or rather AGI~ PBCGP the side of one cuneus , and PAGE the side of another , their common edge being the line PG perpendicular to the base AC ; P being the required point in the side AB where the attraction of the section ABC , or indefinitely thin cuneus , shall be greatest in a direction parallel to the horizon AC .
And then from the foregoing suppositions , it is evident that in whatever point of AB the attraction of ABC is greatest , there also will the attraction of the whole hill be the greatest .
8 .
Now draw HPDEF parallel to AC ; and AH , PG , BI , CF , perpendicular to the same .
Then it is evident that at the point P , in the direction PF , the attraction of PBCGP is affirmative , and that of PAGE negative .
But PBCGP is = PBD + BDE + PFCG EFC ; and PAGE = PHAG PHA .
Therefore the attractions of PBD , BDE , PFCG , PHA , are affirmative ; and those of EFC , PHAG , negative .
Put now BI = a , AI = b , IC = c , AB = d , BC = e , AC =g= b+ c , and PG = x , the altitude of the point P above the bottom .
Also let s= the sine of the indefinitely small angle of the cuneus to rad .
1 ; and q2 =~ a2 g2 2abgx + d2 x2 .
Then by Art .
3 , the attraction PBD is s. PD. BD/ BP = sb Xa x/ d , of PHA is s. PH .
PG/ PA = sb X x/ d. By Art .
4 , the attraction PFCG is s. PG X h. l. PF + PC/ PG = sx X h. l. ag bx + qq/ ax , of PGAH is s. PG X h. l. PH + PA/ PG = sx X h. l. b+ d/ a. By Art .
5 , the attraction of EFC is s. OF .
FC/ EC2 X PC PE PE .
OF / EC X h. l. EC2 + EC .
PC + PE .
OF / PE .
EC + PE .
OF = sc/ ee X qq g. ax gc .
a x/ e X h. l. acg + eqq + aax bcx/ g. c+ e. a x. Lastly by Art .
6 , the attraction of BDE is s. BD .
DE/ BE2 X PB PE + PE .
DE/ BE X h. l. BE2 + BE .
BP PE .
DE/ PE .
BE PE .
DE = sc/ ee .
axXdg+ cg/ e X h. l. ee + de cg/ eg cg .
These quantities being collected together with their proper signs , and contracted , we have ab/ d + C. ad qq dx/ ee +xX hyp .
log .
ag + qq bx/ b + d. x+ sX ccg .
a x/ ~3 X h. l. ee + de cg .
acg + eqq + a2x bcx/ gg .
ee cc .
ax for the whole attraction in the direction PE .
9 .
Having now obtained a general formula for the measure of the attraction in any sort of triangle , if the particular values of the letters be substituted which any practical case may require , and the fluxion of this attraction be put = 0 , the root of the resulting equation will be the required height from the bottom of the hill .
10 .
But for a more particular solution in simpler terms , let us suppose the triangle ABC to be isosceles , in which case we shall have d= e , and g= 2b = 2c , and then the above general formula will become 2ad qq dx/ dd b+xX h. l. 2ab + qq bx/ b + d. x sX +a x/ d3 .
2b3 X h. l. 2ab2 + dq2 b2 a2 .
x/ 2b2 .
ax for the value of the attraction in the case of the isosceles triangle , where q2 is =~ 4a2b2 4ab2x + d2x2 .
And the fluxion of this expression being equated to 0 , the equation will give the relation between a and x for any values of b and d , by a process not very troublesome .
11 .
Now it is probable that the relation between a and x , when the attraction is greatest , will vary with the various relations between b and d , or between b and a. Let us therefore find the limits of that relation , between which it may always be taken , by using two particular extreme cases , the one in which the hill is very steep , and the other in which it is very flat , or a very small in respect of b or d. 12 .
And first let us suppose the triangular section to be equilateral ; in which case the angle of elevation is 60~ , which being a degree of steepness that can scarcely ever happen , this may be accounted the first extreme case .
Here then we shall have d= 2b = 2/ 3a~3 , and the formula in Art .
10 , will become s X : 2a r x/ 2 +xX h. l. 2a + 2r x/ 3x +a x/ 4 X h. l. a+ 2r + x/ a x for the value of the attraction in the case of the equilateral triangle , in which r is =~ a2 ax + x2 .
13 .
Or if we take x= na , where n expresses what part of a is denoted by x , the last formula will become sa X : 1 1/ 2n 1/ 2 ~1n+ n2 +nX h. l. 2n+2~1n+ n2/ 3n + 1 n/ 4 X h. l. 1+n+2~1n+ n2/ 1 n for the case of the equilateral triangle .
14 .
To find the maximum of the expression in the last article , put its fluxion = 0 , and there will result this equation 1+1+ n/ ~ 1n+ n2 =2 h. l. 2n+2~1n+ n2/ 3n 1/ 2 h. l. 1+n+2~1n+ n2/ 1 n ; the root of which is n =.251999 .
Which shows that , in the equilateral triangle , the height from the bottom to the point of greatest attraction , is only 1/ 500th part more than 1/ 4 of the whole altitude of the triangle .
And this is the limit for the steepest kind of hills .
15 .
Let us find now the particular values of the measure of attraction arising by taking certain values of n varying by some small difference , in order to discover what part of the greatest attraction is wanting by observing at different altitudes .
16 .
And first using the value of n ( .251999 ) as found in the 14th article , the general formula in Art .
13 , gives sa X 1.0763700 for the measure of the greatest attraction .
17 .
If n= 3/ 10 , or x= 3/ 10 a ; the same formula gives sa/ 20 X : 17 ~ 79 +6 h. l. 17 +2~ 79/ 9 + 7/ 2 h. l. 13 +2~ 79/ 7 = sa X 1.0702512 for the attraction at 3/ 10 of the altitude , which is something less than the other .
18 .
If n= 4/ 10 = 2/ 5 ; the formula gives sa/ 20 X : 16 ~ 76 +8 h. l. 8+~ 76/ 6 +3 h. l. 7+~ 76/ 3 = sa X 1.0224232 for the attraction at 4/ 10 or 2/ 5 of the altitude ; less again than the last was .
19 .
If n= 5/ 10 = 1/ 2 ; the formula gives 1/ 4 sa X : 3~32 h. l. 3+ 5/ 2 h. l. 3+2~3= sa X.9340963 for the attraction at half way up the hill ; still less again than the last .
20 .
If n= 6/ 10 = 3/ 5 ; the formula gives sa/ 20 X : 14 ~ 76 + 12 h. l. 7+~ 76/ 9 +2 h. l. 8+ ~76/ 2 = sa X.8109843 for the attraction at 6/ 10 or 3/ 5 of the altitude from the bottom ; being still less than the last was .
And thus the quantity of attraction is continually less and less the higher we ascend up the hill above the.251999 part , or in round numbers.252 part of the altitude .
Let us now descend , by trying the numbers below.252 ; and first , 21 .
If n =.25 = 1/ 4 ; the same formula in Art .
13 , gives 1/ 8 sa X : 7~ 13 +2 h. l. 7+2~ 13/ 3 + 3/ 2 h. l. 5+2~ 13/ 3 = sa X 1.0763589 for the attraction at 1/ 4 of the altitude ; and is very little less than the maximum .
22 .
If n= 2/ 10 = 1/ 5 ; the formula gives 1/ 10 sa X : 9~ 21 +2 h. l. 9+2~ 21/ 3 +2 h. l. 3+~ 21/ 2 = 1/ 10 sa X : 9~ 21 +2 h. l. 23 +5~ 21/ 2 = sa X 1.0684622 for the attraction at 2/ 10 or 1/ 5 of the altitude ; and is something less than at 1/ 4 of the altitude .
23 .
If n= 1/ 10 ; the formula gives sa/ 20 X : 19 ~ 91 + 2 h. l. 19 +2~ 91/ 3 + 9/ 2 h. l. 11 .
+2~ 91/ 9 = sa X.9986188 for the attraction at 1/ 10 of the altitude ; still less than the last was .
And , lastly , 24 .
If n= 0 , or the point be at the bottom of the hill ; the formula gives 1/ 4 sa x2+ h. l. 3= sa x ~7746531 for the attraction at the bottom of the hill ; which is between 2/ 3 and 3/ 4 of the greatest attraction , being something greater than 2/ 3 but less than 3/ 4 of it .
25 .
The annexed table exhibits a summary of the 6/ 10 8109843 1/ 4 calculations made in the preceding 5/ 10 9340963 2/ 13 articles ; where the first column 4/ 10 10224232 1/ 20 shows at what part of the altitude 3/ 10 10702512 1/ 180 of the hill the observation is made ; 252/ 1000 10763700 0 the second column contains the 1/ 4 10763589 1/ 97852 2/ 10 10684622 1/ 134 corresponding numbers which are 1/ 10 9986188 1/ 14 proportional to the attraction ; and 0 7746531 2/ 7 the third column shows what part of the greatest attraction is lost at each respective place of observation , or how much each is less than the greatest .
26 .
Having now so fully illustrated the case of the first extreme , or limit , let us search what is the limit for the other extreme , that is , when the hill is very low or flat .
In this case b is nearly equal to d , and they are both very great in respect of a ; consequently the formula for the attraction in Art .
10 , will become barely s x : xx h. l. 2a x/ x + 2 .
axx h. l. 2a x/ a x ; the fluxion of which being put = 0 , we obtain 0= h. l. 2a x/ x 2 h. l. 2a x/ a x= h. l. 2a x/ x h. l. 2a x/ a x~2 = h. l. a ~~2/ x. 2a x ; hence therefore a x~2 = x. 2a x , and x=ax1 ~1/ 2 = ~2929a .
Which shows that the other limit is 29/ 100 ; that is , when the hill is extremely low , the point of greatest attraction is at 29/ 100 of the altitude , like as it is at 25/ 100 when the hill is very steep .
And between these limits it is always found , it being nearer to the one or the other of them , as the hill is flatter or steeper .
27 .
Thus then we find that at 1/ 4 of the altitude , or very little more , is the best place for observation , to have the greatest attraction from a hill in the form of a triangular prism of an indefinite length .
But when its length is limited , the point of greatest attraction will descend a little lower ; and the shorter the hill is , the lower will that point descend .
For the same reason , all pyramidal hills have their place of greatest attraction a little below that above determined .
But if the hill have a considerable space flat at the top , after the manner of a frustum , then the said point will be a little higher than as above found .
Commonly , however , 1/ 4 of the altitude may be used for the best place of observation , as the point of greatest attraction will seldom differ sensibly from that place .
And when uncommon circumstances may produce a difference too great to be entirely neglected , the observer must exercise his judgment in guessing at the necessary change he ought to make in the place of observation , so as to obtain the best effect which the concomitant circumstances will admit of .

