Volume 47, 1914
This text is also available in PDF
(317 KB) Opens in new window
– 578 –

### Art. LIII.—On the Deflection of the Plumb-line from the Vertical due to the Spheroidal Form of the Figure of the Earth, and its Effect on the Latitude of Stations above the Mean Sea-level.

[Read before the Otago Institute, 22nd September, 1914]

It is now well known that the attraction of mountains and elevated regions has a sensible influence on the plumb-line. These local deflections have been the subject of numerous investigations since 1854, when Archdeacon Pratt computed the effect of the attraction of the Himalaya Mountains on the plumb-line in North India (Phil. Trans., 1854).

In modern geodetic surveys the degree of precision in the determination of the apparent astronomical latitude by means of the zenith telescope is so great that the station error, as it is called, or the difference between the observed and computed latitudes, due to local causes, has become a part of the routine work.

The deflection of the plumb-line by the attraction of mountains or by the variation of the density of the underlying strata will not be further discussed, the object of the present investigation being to determine the

– 579 –

amount of deflection and its direction that a plummet will deviate from the vertical line, or normal, at its point of suspension, in the plane of the meridian, for any length of the suspending cord, due to the spheroidal form of the figure of the earth.

### Statics of the Problem of a Plummet suspended from a Fixed Point.

Let PEP1E1, fig. 1, represent the meridian ellipse of the earth, PP1 the axis of rotation or polar axis, and EE1 the equatorial axis.

We can infer from consideration of the symmetry of the figure that a plumb-line suspended at the Poles or at the Equator does not deviate from the vertical, and that at all other latitudes there is a deviation.

Let A be a point at the mean level of the sea, the latitude of which is φ', and suppose p a plummet suspended from the point A. Such a suspension can be actually obtained by means of a mine-shaft for small depths, but for the statement of the problem the existence of an opening in the earth need not be considered, for the solution of the problem gives the direction that a shaft takes when sunk by means of the plumb-line.

Let C be the centre of the earth, F and F1 the foci.

The forces acting on p are the attraction of the earth and the tension of the suspending cord, the weight of the cord being neglected.

Through p let a confocal ellipse be described, and let it generate a spheroid by revolving about the axis PP1. It is a well-known theorem in attractions that the potential within a spherical or ellipsoidal shell is constant, and therefore the force acting on p is the attraction of the above-generated spheroid, the surface of which is equipotential (Routh's “Analytical Statics,” ii, p. 104).

A second theorem in attraction is: If there be any resultant force, due to any attracting mass, this force acts along the normal to that equipotential surface on which the point lies (Pierce's “Newtonian Potential Function,” p. 38).

Since the tension of the cord is the resultant force, it is normal to the confocal ellipse, and therefore bisects the angle Fp F1.

Thus for all positions of p the direction of the resultant is towards the point A, and the angle FpF1 is bisected by Ap produced.

– 580 –

### Geometry of the Problem.

The problem to be solved can be stated as follows, in its most general form: Given the three sides of a triangle ABC, it is required to find the locus of a point, such that the line from the angle A to the point bisects the angle included between the lines from the angles B and C to the point.

The bisected angle contains the point A, and no restriction is placed on the size of the angle except that it is less than four right angles.

The deviation of the plumb-line is a particular case of the general form.

Let ABC, fig. 2, be the given triangle, and D a point such that the angles ADB, ADC are equal.

Take A as the origin of co-ordinates and the line bisecting the angle BAC as the axis of X.

Let r, θ be the polar co-ordinate of D. Denote the equal angles ADB, ADC by &chi;.

To obtain the equation to the locus of D, we have from the triangle ABD

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

r/C=sin(A/2+θ-&chi;)/sing&chi; ∴cot&chi;=r+c cos (A/2+θ)/c sin (A/2+θ) (1)

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Again, from the triangle ACD we find cot&chi;=r + b cos (A/2-θ)/b sin (A/2-θ) (2)

– 581 –

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Eliminating cot &chi; we obtain sin 2θ-mr sing θ+nr cos θ=0 (3)

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Wherein m has the value b+c/bc cos A/2, and n the value b-c/bc sin A/2, or 2 sin θ - mr sin θ + nr = o, since θ is small. ∴ sign θ=2nr/2-mr (4)

The amount of the deflection is r sin θ, and is towards the Equator, since θ is positive.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

To transform the equation to rectangular co-ordinates we substitute in (3) the values of the variable in terms of x = r cos θ, y = r sin θ, and x2 + y2 = r2, and we have nx3-myx2+2xy+nsy2-my3 = 0 (5)

When B and C are the foci of an ellipse the triangle ABC is isosceles if A is at the extremity of the minor axis.

Then m = 2/b cos A/2 and n = o.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Substituting these values of the constants in equation (5) we obtain 2xy-2/bcos A/2yx2-2/bcoxa/2y3 = 0

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Therefore y = o (6) and y2 + (x-b/2secA/2) = (b/2secA/2)2 (7)

Equation (6), y = o, represents the axis of X, and for a plummet suspended at the Pole of the earth the axis of X is the vertical, and consequently there is no deviation from the vertical when the plummet is suspended from the Pole of the earth considered as an oblate spheroid.

Equation (7) represents a circle whose radius is b/2 sec A/2, and is therefore the circle circumscribing the triangle ABC.

The co-ordinates of the centre are y = o. x = b/2 sec A/2.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

When the point A coincides with the extremity of the major axis of the ellipse the angle A vanishes, and we find m = b+c/bc and n = o. These values of the constants substituted in equation (5) reduce it to the following two results:—y=0 (8) y2+(x-bc/b+c)2 = (bc/b+c)2 (9)

The equation (8) represents the axis of X, and is the solution of the problem when the plummet is suspended at the Equator, from which we infer that for places situated on the Equator there is no deviation of the plumb-line from the vertical.

– 582 –

Equation (9) represents a circle of radiu bc/b+c, the co-ordinates of the centre being y = o. x = bc/b+c

If in fig. 1 the inner ellipse represents the surface of the earth and A a point at an elevation above the surface, from which a plummet is suspended by a cord Ap, the line Ap produced bisects the angle FpF1, and is normal to the surface of the earth at p; and the point A is therefore vertically above p. This condition holds for any length of the suspending cord, and the plumb-line at the earth's surface suspended from a point above it always coincides with the vertical.

The level surface at p is at right angles to Ap, and the angle EHp is the geographic latitude.

The level surface at A is at right angles to the line AK, which bisects the angle FAF1, and the geographic latitude of A is the angle EKA. The difference of these angles is the variation of the latitude of p, due to the spheroidal form of the earth, and the elevation pA.

The level surfaces become more spherical as we recede from the earth, the point K moves along the major axis and ultimately coincides with the centre.

If φ1 denotes the latitude of p and φ2 the angle EKA, then φ1 - φ2 is the angle θ in equation (3), where A is the origin, and the line bisecting FAF1 the initial, or the axis of X.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Let δ0=HAF. δ=AFp. δ′0=HAF1. δ′=AF1p.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

FpF1=ψ. AP=r.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Fp=f1. F1P =f.

It is now proposed to express AF, AF1, and δo, δ'o in terms of pF, pF1, and r.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

The triangle AFp gives AF2f12+r2+2f1r cosψ/2 (10) f1 sin δ = r sing δ0 (11) δ + δ0 = ψ/2 (12)

Similar expressions are obtained from the triangle AF1p for AF1, δ‘o, and δ’.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

From (11) sin δ = r/f1 sin δ0 = p1 sin (ψ/2 - δ), where p1 = r/f1 (13)

For even the highest mountain-tops the values of p1 and δ are small, and equation (13) can be developed by Maclaurin's series, thus: δ = tan 1 (p1sin ψ/2/1+p1 cos ψ/2) = p1 sin ψ/2-½ p12 sin ψ +&c (14)

– 583 –

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Similarly, if p = p=r/f δ′ = p sin ψ/2 - ½ p2 sing ψ + &c (15)

To take a numerical example, suppose that in latitude 45° S. it is required to find, the difference in latitude between the top of a hill, 490 metres high, and the sea-level vertically below it, using Helmert's spheroid.

The value of θ can be computed by equation (3), but it can be found directly from (12), (14), and (15), as under.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

ψ/2 = δ +δ0 = δ′+δ′0 δ′00 = δ -δ′ (16) Now, δ′0 = A/2 +θ. δ0 = A/2-θ. ∴δ′00 = 2θ. or θ ½ (δ′0 - δ0) (17)

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

And by the aid of (14) and (15) θ in seconds of arc is θ″=½ p1-p/sin 1″ sin ψ/2 - ¼. p12-p3/sin 1″ sing ψ +&c (18) p1-p = r/f1 - r/f = r (f-f1)/f f1 = 10 -6x 893.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Computing the first term of (18) from the known value of sin ψ/2 we find θ =0″.053

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Thus the latitude of the hilltop is φ2= φ1 - θ. = 44° 59′ 59″.947.

These differences of latitude are so small that in most cases they will be masked by the greater differences due to local conditions, which deflect the plumb-line; they possess the advantage that they can be easily computed, and the observed value can be corrected or reduced to the sea-level.

A summary of the three principal results is as follows:—

 (1.) A plummet suspended from a point at the sea-level deviates from the vertical in the plane of the meridian, except at the Poles or the Equator; the amount of the deviation depends on the length of the suspending cord, and the deviation is towards the Equator. (2.) A plummet suspended from a point above the surface of the sea-level, when the suspending cord just reaches the surface, always coincides with the vertical for any length of the suspending cord. (3.) At heights above the surface the observed latitude is less than its corresponding value at sea-level.