Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Parasnis - 1951 - Study Rock Midlands

Download as pdf or txt
Download as pdf or txt
You are on page 1of 20

A STUDY O F ROCK DENSITIES IN THE ENGLISH MIDLANDS

D . S. Parasnis

(Communicated by A. H. Cook)

(Received 1951 June 14)

Summary
A knowledge of the densities of rock formations is important in the inter-
pretation of gravity anomalies. This paper describes the reaults of field and
laboratory experiments made on the rocks in the Midlands of England.
Density measurements were made on nine geological formations from the
Silurian to the Cretaceous and also on Metamorphic rocks from the Malvern
area. A catalogue of densities has been prepared for general use based on the
assumption that rocks in the field are saturated with water. A table of
densities and porosities for a large number of rocks is also given. Among the
conclusions drawn about the observed sampling variances, the most significant
is that rock formations in the West Midlands do not appear to show any
'' regional variation " of density. The results of the field and laborator)
methods are consistent within the experimental errors of each, in marked
contrast to the results reported by S. Hammer from observations in some
American mines, from somewhatdifferent data.

Introduction
I.
I. I. The Bouguer gravity anomalyat any place is defined as the observedvalue of
gravity minus the theoretical value as calculated from the following well-known
formula :
+
gdc.=yo - klh k,ph - T.
Here yo is the value of gravity as given by the International Gravity Formula at the
latitude of the station, Klh is the so-called " free air " correction due to the height (h)
of the station above sea-level, and Ksph is the attraction of an infinite slab of material
of density p supposed to lie between the station and the sea-level. The assumption
that this material is in the form of an infinite slab may not be justified in regions
where the topography surrounding the station is not flat and the term ( - )'2 is
included to allow for the attraction of the topography. T is always negative
whether there is a hill near the station or a hollow. The values of the constants kl
and K, when h is expressed in feet (as it is, throughout this paper) are given by:
k, =0.09406mgal/ft.
kap =0*0128p mgal/ft.
It isevident that to calculate the Bouguer anomaly it is necessary to know the density
of the geological formation underlying a station. An error of o.lg/cma in the
density corresponds to an error of nearly 0.13 mgal in the Bouguer anomaly for
every IOO ft. This in itself is not a very large error but an error of 0.1g/cm8in the
density can have a very large effect on the interpretation of the anomalies because
we may be concerned with two adjoining rock formations whose densities may

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
A study of rock densities in the English Midlands 253
differ by 0.3 or o-4g/cmaonly. In the case of the “step-structure” not un-
commonly encountered, not only will the depth be grossly in error but also the
form of the step might be inverted. It is clear, therefore, that density measure-
ments on rocks are of fundamental importance in gravity surveys. In this paper an
account is given of an investigation to determine the densities of rocks by two
independent methods. In addition, measurements in mines made by Cook and
Thirlaway (1951) are discussed.
1.2. The field method-Nettleton (1939) has described a method of density
measurement. He calculates gravity profiles for various assumed values of the
density of the underlying formation and selects that one which makes the profile
least correlated with the topographic section. An advantage of this method is that
the regional trend of gravity, if any, needs no separate attention, but the dis-
advantage is that a rather large number of calculations is required, and at best two
values of the density are obtained giving nearly the same result between which the
real value is presumed to lie. Further, in this method an estimate of the standard
deviations can be made only with a laborious arithmetical procedure.
The author’s method is slightly different from Nettleton’s method and is
described below. Suppose that there are two stations at which the observed
values of gravity areg, andg,. Letg; andgi be the respective calculated values of
gravity. Then
g; =Yo2 - kh2 + k2Ph2,
g; =Yo1 - Wl + hPh1,
where &, h, are the heights of the stations above sea-level. Now
g2=gL + Bouguer anomaly at station 2,
g, =g; + Bouguer anomaly at station I.
If the Bouguer anomalies at the stations are the same,
g2 -g1 =A-g; ‘Yo2 -Yo1 - h(h2 - 4) + k2@2 - hl)P (1)
g2-gl can be observed by a gravimeter and as all other quantities except p can be
found, we have an equation to determine p by the least squares method. In this,
of course, it is assumed that the Bouguer anomaly is known at the two stations.
But actually it cannot be known without the knowledge of p. Hence, in practice
we assume the density obtained from measurements on rock samples to find the
trend of the anomaly.
If the apparent topographic corrections T,, Tl at the two stations are appreciable
they can be found by using tables such as those of Hammer (1939) which are usually
calculated for an assumed density of po. Then the corrections for a density p are
T2p/po and T$/po respectively. If these are included in equation (I), we have
Ag =gz -g1= Yo2 -Yo1 -kl(h2 - hl) + [kz(h, -hl) - (T*- Tl)/POlP
=4 0 - klAh + [ W h - (T2 - Tl)/POlP
or
Ag - AYO + kl& = [ka& - (Tz- ~ l ) / P O l P . (2)

Let us denote the left-hand side by Ay and the right-hand side by PAX. Now, if
a line of stations at different heights be established, then taking one of them as a
base, we can calculate Ay and Ax for each of the stations. The graph of Ay against
Az will be a straight line (through the origin) and its slope will be p. In the next
section such measurements made over a number of formations are described.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
254 D. S. Paramis
The simplest way to apply this method is to make the measurements at various
points on the slope of a hill. Special attention must be paid to the following
factors :
(I) The hill must be an erosion feature and not an anticline or a faulted
structure. River valleys often provide excellent examples of such features.
(2) The surrounding gravity field must be not abnormally disturbed by
structures such as intrusions, troughs, faults, etc., otherwise a fictitious
density will be obtained.
(3) The hill must be of a sufficient height so that the differences of gravity
between consecutive stations are appreciably greater than the standard
deviation of a single determination.
(4)The topography of the surrounding country must be allowed for.
1.3. The laboratory method.-In this method samples of rocks are collected
from outcrops near the station and their density is found from laboratory deter-
minations. This method, however, may have the following drawbacks because
the outcrop may not be truly representative of the strata below :
(I) If the strata consist of bands of more than one lithological character (such
as alternating bands of shales and limestones) the density from a specimen
of one character may be widely different from that found from another.
The “ effective ” density determining the Bouguer anomaly will not simply
be a mean of all such densities but will depend on the frequency ofeach band
within the formation.
( 2 ) The rocks may be compacted at great depths. Even with specimens of the
same lithological character, the density may be different from different
horizons. Thus the density of the Oxford clay (Athleta zone) in the
Stewartby brickwork pits was found to vary from 1-85to 1-88from the
surface to a depth of about IOO ft.
(3) It is usually not known with certainty to what extent the rocks contain water
and the best assumption is that the “true” density lies somewhere between
the density of a saturated rock specimen and that of a dry one.
(4) The weathering of rocks may affect their density considerably.
The field method is the more basic of the two methods but also the more
difficult one. The laboratory method is easy but not necessarily correct owing to
the causes just mentioned.

The fild measurements of density


2.
2.1. The measurements were made with an Askania type gravimeter belonging
to the Department of Geodesy and Geophysics, Cambridge University. A dis-
cussion of the accuracy of measurements with this instrument has been made by
Cook (1951). The smallest division on the scale corresponds to 0.06 mgal but
the final accuracy of measurement is usually about 0.18 mgal, owing chiefly to the
drift of the scale. The calibration has a standard deviation of 0.42 per cent.
Some measurements by the field method have been made previously by Cook and
Thirlaway (1951) and, using Nettleton’s method, by the Anglo-Iranian Oil
Company in their gravity surveys (privately circulated work). But both these
measurements were restricted in the number of the lithological formations
covered. In the present investigation an extended study of the rock densities in

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
A study of rock densities in the English Midlands255
the West Midlands has been made. In the next few sections the following
measurements are described :
Upper Lias at Upton St. Leonards (Glos.).
Silurian Limestone and Shales at Woolhope Cockshoot (Herefords.).
Old Red Sandstone at Hanley William (Worcs.).
Old Red Sandstone on May Hill (near Monmouth).
Middle Coal Measures Shales and Sandstones-Forest of Wyre (Worcs.).
Carboniferous Limestone-Forest of Dean (Glos.).
Triassic Lower Mottled Sandstone-Bridgnorth Cliffs (Shropshire).
780 r Inferior Oolite
2304 collected here I

Middle Lias

h e r Lias
2000 6000 loo00 14Ooo ft.
I.-Profile of the hill near Upton St. Leonard zuith locations of gravimeter stations.
(Section along line AA' on Fig. w.)
Ordinates: Height in feet aBwe O.D.

AY
mgal

FIG.2.-Density line at Upton St. Leonard.


The observation on Lias is treated first because it illustrates some important
points concerning the hill measurements.
2.21. Upper Lias at Upton S t . Leonards (Gloucestershire), Lat. 51" 49' 51" N.,
Long. zo12' 11" W.* Fig. I shows the simplified geological structure of the hill
* The latitude and longitude refer to the lowest station on the particular hill.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
256 D. S. Parasnis
which is the erosion work of numerous small streams. The observations, which
are reproduced in detail in Table I, were made along the road from Upton to
Painswick (B4073). Although the hill is made of three distinct layers of rocks,
the observations indicate only one straight line, which means that the densities of
these layers are not very different from each other. Fig. 2 is a graph of Ay against
Ax.
TABUI
Observations near Upton St. Leonards
Base 2300. Height of Base above O.D.=183 ft. T1=0.22mgal. p0=2 g/cms.

I I1 111 IV V VI VII VIII IX x

2307 - 10.62 -1.5 + 162 +I5'24 2'074 0.35 2.06 6-12


2301 - 18.62 -1.8 +274 25'77 3-51 0.48 3.38 8-95
2302 -20.42 -0.9 +296.4 27-88 3'79 0.35 3'73 8.36
2303 -26.24 - I '4 +402.2 37'83 5-15 0.40 5a 0 6 I 3 '00
2304 -37.38 -1.5 +539 50.70 6-90 0.48 6.77 14-82
2305 -28.54 -3.4 +398*5 37.48 5-10 0.23 5.10 12-34
2306 -32.26 -3.1 +454'7 42'77 5'82 0.38 5 '74 13-61
2300 0 0 0 0 0 0'22 0 0
(Base)

As 2-5
mgal 2.0
-. 2303
+
1.5 ' 'Yl
+2301
2&5 -
1.0
+0.5
.- 2304 + 2306
+ I
2 p 2 o p o 4opo bpoo eOp0 lap00 Uapo 14,900 16,900
0 +
-0.5 . 2302

1.0 -
1-5 -
2.0 -
2.5 b

The best straight line is found to be


Ay = (2.28 f o*I~)Ax+ (0.574 & 0.83).
The place where samples were collected is shown in Fig. L+A which is a simplified
map of the region. The regional trend in this as well as all the other cases was
found to be insignificant. The examination is made by plotting the residua1
-PAX (mgal) against D, the distance in feet from the base assuming
As = AyobS.
AS =K'D + b.
As an example, the graph of As against D in the present case is shown in Fig. 3.
The best straight line is found to be
As =(0-47& 0.80) x 1o-*D+ (0.73 0.62) mgal.
This shows that the value of k' is not significantly different from zero.
Result fm density:
p =2-28 & 0.19 g/cm8.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
A study of rock densities in the English Midlands 257

0 lo00 2o00ft.
FIG.*.-Area near Upton St. Leonards ( U p p e Lias).
~ Based on the O.S. Map, with the sanction
of the Cont~ollerof H.M. Stationery O@e.
Geological section along AA' is shown in Fig. I.
0-

0
+
Contours of topography at 100 ft.

Exposuresfrom which samples were taken.

Gavimeter station.
interval.

1J Key for all


the maps.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
258 D. S. Parasnis
2.22. Silurian limestone and shales: Ludlow beds at Woohope Cockshoot (Here-
fordshire). Lat. 52' 02' 03"N.,Long. 2' 31' 1 9W. ~
Site : The hill has been eroded by the tributaries of the Preston Brook in the
Woolhope anticline. It is composed of alternating bands of limestone and shales,
the dip being slightly to the south-east.
Samples: The samples were collected in the quarry just above Woolhope
Cockshoot. The quarry is marked in Fig. 4~ which is a map of the area.
G a v i t y measurements : Measurements were made at five stations. The height
of the lowest one above O.D. is 305 ft. and that of the highest one 734ft. The best
straight line obtained is
Ay = (2.73k 0-22)Ax+ (0-109 k 0.63).
The topographic correction was of the order of 0-5mgal and has been applied.
Result for density:
p = 2-73f 0.22 g/cm3.

I
0 lo00 ft.
FIG.q~.-Ar ea near Woolhope Cockshoot (Silurian limestone and shales). Based on the O.S.
M ap, m'th the sanction of the Controller of H.M. Stationery O m e .
2.23. Old Red Sandstone at Hanley William ( Worcestershire). Lat.
52O 18' 43"N.,Long. 2' 29' 43"W.
Site: This hill, which is eroded by the tributaries of the Teme, is made of
Downtonian rocks (Lower Old Red Sandstone). Wickham King (1925)mentions
that along the lower Teme valley the strata consist of the Middle and Lower
Downtonian. The hill is nearer the lower part of the Teme and it is estimated that
it is composed of about ZOO ft. of red and green marls underlain by about 300ft. of
calcareous and marly sandstone. Below this is a purple-green fine sandstone
go ft. in thickness underneath which is a thickness of 400-500ft. of " deep compact
marls devoid of sandstone". This is followed by Lingula marls (about 315 ft.),
Ledbury group (400ft.) and Temeside group (about 150ft.).
Samples: These were collected from an exposure by the Long Bank that is
marked on Fig. 412.
Gravity measurements: Measurements were made at nine points on the hill.
The height above O.D. of the lowest station is zooft. and that of the highest one
so0 ft.
The best straight line is
Ay =(2-47k 0.048)Ax- (0.18& 0.23).
The topographic correction was found to be negligible.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
A study of rock densities in the English Midlands 259

P I a I I
I
0 ‘lo00 2000 3000 4Ooo 5o00ft.
FIG.qc.-Ma@ of the region near Hanky Willah (Old Red Sandstone). Based on t h O.S. Map,
with the sanction of the Controller of H.M. Stationery Ofice.
G 10

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
260 D . S. Parasnis
Result for density :
p = 2.47 I0.048 g/cms.
2.24. Old Red Sandtone at May Hill near Monmouth (Ditton Series). Lat.
51" 48' 40"N., Long. '
2 42' 02"W.
Site: The hill has been eroded chiefly by the River Wye and is composed of
Lower Old Red Sandstone beds.
Samples : These were collected from an exposure near the " Fiddler's Elbow "
marked on Fig. @.

-
0 IOoOft.
FIG. q~.-May Hill near Momnauth (OM Red Sandstone). Based on the O.S. Map, with the
sanction of the Controller of H.M. S t a t i w Ofice.

Gravity measurements: The measurements were made at six stations along


"Staunton road" which goes over the hill from Monmouth to Staunton. The
height above O.D. of the lowest station is 77 ft. and that of the highest one 551ft.
The topographic correction was found to be as high as I-15mgal for Station
No. 2309 ; for others it was of the order of 0.50 mgal. The best straight line is
Ay =(2.46f 0.08)Ax + (0.16& 0.31).
Result for density :
p = 2.46 f 0.08 glcms.
This compares well with that at Hanley William in the previous section.
2.25. Carboniferous Limestone in the Forest of Dean (Gbucestershire). Lat.
50" 51' 00"N., Long. 2 ' 35' 40"W.
Site: The hill is eroded by the River Wye and its tributaries and is situated
outside the north-west margin of the Forest of Dean near Lower Lydbrook in the
Carboniferous Limestone strip that girdles the coalfield area.
Samples : These were collected from a quarry near Pope's Grove towards the
northern part of the hill. Fig. 4E is a map of the region.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
A study of rock densities in the English Midlands 261
Gravity rneasurenents : Eight stationswere available. The height above O.D.
of the lowest one is 147-1ft. arid that of the highest one 635-5ft. The best straight
line is
Ay =(2.54 k 0 - 0 7 ) h- (0.25 & 0.11).
The topographic correction is negligible.
Result for density :
p = 2.54 f 0.07g/cm3.

c a a I
0 lo00 2000 3o00ft.
FIG.qB.-Fwest of Dean near Lower Lydbrook (CarbomferouzLimestone). Based on the O.S.Map
with the sanction of the ControUer of H.M. Stationery O f i e .
G Is*

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
262 D . S. Paramis
2.26. Middle Coal Measures in the Forest of Wyre (Worcestershire). Lat.
52O 23' IO'N., Long. 2
' 24' 40"W.
Site : The hill is situated near Far Forest and has been eroded by the Dowles,
Lem and Baveney brooks. The exposures in the Dowles valley belong to the
Middle Coal Measures down to about 880 ft. from the surface (Kidston, Cantrill
and Dixon, 1917). They do not here bear any workable coal. A little farther
north (at Billingseley), however, it is known that the sweet coal group is present.
Samples: These were collected from an exposure near the confluence of the
three brooks north of the Quarry cottages. Fig. 4~ is a simplified map of the
region.
Gravity measurements : Nine stations were available for these. The height of
the lowest one above O.D. is 255 ft. and that of the highest one 468ft. The best
straight line is
Ay = (2.70f 0.1o)Ax- (0.039 f 0.19).
The topographic correction is only of the order of 0.1mgal.
Result for density:
p = 2-70& 0.10g/cms.
2.27. Trimsic Sandstone of Brienorth Clays (Shropshire). Lat. 52' 32' 18" N.,
I,ong. 2O 24' 42"W.
Site : Bridgnorth is situated on the Castle Hill (height about 150ft.) eroded out
of Lower Mottled Sandstone by the River Severn. One observation was made at
the church on the top of the hill and the other on the bank of the river. It was not
found practicable to make observations other than these two.
Result for density:
p = 2-30& 0.13g/cm8.

The topographic correctionwas found to be 0-2mgal.


Samples : These were collected at various places on the road from Bridgnorth
to Quattford. There are several good exposures.

3. Laboratory measurements of density


3.1. One object of these measurements was to obtain independent values of
rock densities for comparison with the gravimeter values. A second object was
to investigate the sampling variances of the densities when several specimens of the
same lithological character are obtained from the same exposure and alsowhen they
are obtained from different exposures. Samples were collected from the hills on
which gravimeter measurements were made and the exact position of the exposure
was marked on the &inch Ordnance Survey map for future reference. A com-
parison of the densities obtained by the two methods would show whether a random
sampling is sufficient to determine the density to be used in the calculation of
anomalies.
3.2. The measurement of density.-It has been pointed out that the " field"
density of a rock must lie between the saturated and the dry density. The follow-
ing method was used to saturate the rock specimen:
The specimen was placed under water in a small vessel connected to a vacuum
pump. The vessel was evacuated down to about ~ o m mof mercury and the
specimen kept under vacuum for about 30 minutes after air bubbles had ceased to
come out of it. It was then left under atmospheric pressure for about 16 hours.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
A study of rock densities in the English Midlands 263

FIG.@.-Part
-
0 loo0 2Ooo ft.
of the Forest of Wyre (Middle Coal Meusures). Based on the O.S. Map, with the
sanction of the Controller of H.M. Stationmy m e .

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
264 D. S. Parasnis
This method is slightly different from that used by Thirlaway (see Cook and
Thirlaway, 1951). The following example shows the difference in weights when
the specimen was weighed at various instants during the process of saturation.
Specimen-Chalk
Weight immediately after evacuation for 30 minutes ... 34-27g.
Weight after 6 hours’ soaking ... ... ... ... 34.38g.
Weight after 16 hours’ soaking ... ... ... ... 345Ig.
Weight after 30 hours’ soaking ... ... ... ... 3451g.
It was found convenient to leave the specimen soaking overnight and this
usually meant 16 hours’ soaking.
The specimen was dried by heating in an oven at IIO deg. C for about three
hours. This time was found sufficient to dry it completely. It was cooled in a
desiccator.
The following procedure of measurements was adopted :
Filter paper was used to remove the surface moisture from the saturated
specimen which was then weighed in air and water. It was next dried by heating
and finally weighed again in air.
Let V be the volume occupied by the grains in a specimen and let o be the total
volume of the pores. Let p o be the grain density, p, the density when saturated
and pa that when dry.
If w, =weight of the saturated specimen in water,
w, =weight of the dry specimen in air,
w3=weight of the saturated specimen in air,
it follows that
.. ..
wl = V(p,- I). ( a ) ; w, = Vp,. (b) ; and w3= Vpo+ w (c). ...
Further, p,=wa/(V+o) and pd=w,/(V+w),
because the total volume is V + w.
From (u) and (b) we obtain
Po =W d W 2 - W l ) ,
and hence also
Po =P d f{ P d - W l / ( V + 4 1 * * (4,
pn can be expressed as I + wl( V + w).
Therefore ( d ) becomes
PO = P d / ( P d - P n + 1) =P~/CI -A)* where A = p a - P*
Also, v/V which is called the void ratio and is denoted by E , can be seen to be
equal to
A/(I-A).
We can define the porosity of a specimen as xooA/(x -A).
Ordinary tap water was used instead of distilled water in all the measurements
but no detectable difference was found in the weights, as the following example
shows :
Specimen-Granite (Dartmoor)
Weight in air ... ... ... ... 88-073g.
Weight in tap water ... ... ... 54060g.
Weight in distilled water ... ... 54060g.
Density of specimen ... ... ... 2-58g/cms.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
A study of rock dmsities in the English Midlands 265
Allowance must be made for the temperature of water. This is done by divid-
ing the density obtained by the density of water at the temperature concerned.
This allowanceamounts to I part in 1000at temperatures of 15-20 deg. C and hence
is quite negligible. Where this was not the case in the following measurementsthe
necessary correction has been made.
3.3. As data on density measurements of rocks are not very plentiful it is
thought desirable to reproduce the results in detail. The existing information in
Great Britain is more or less confined to a paper by Moore (1902), another one by
Holmes (1921) and the work of Cook and Thirlaway (1951). Table I1 shows the
results and the columns are self-explanatory. The standard deviation of a single
measurement of density is +o.o~g/cm*, whence it follows that the standard
deviation UA of A is k 0.014g/cm*. These were not found to vary with lithology.
The standard deviation usof E is equal to [I/(I -A)a]uA and Table I11 gives some
values of I -A and a,.
The maximum error in E in the observed values of Table I1 is of the order of
& 0.06 per cent. The standard error in the value of pu is k 0.05 g/cma for p g = 3
and I -A = 0.9. The error, which is inversely proportional to ( I - A)z, increases
with p,, and as a value of pg = 3-00 g/cm* is rather a high one, f 0.05 g/cm* may be
taken to be the maximum standard error of pr
3.4. The total variance in the measurement of rock density is the combined
effect of 4, u; and 6 where these are respectively the contributions of the errors of
measurement, the variation within one exposure and the regional variation of a
certain type of rock. The important question is whether the variance of density
within one exposure can also account for the variance among different exposures
of the same rock. If there are k samples in each exposure, then the variance of a
single measurement is 4 + 4,while the variance of the mean of these k samples will
be (%+4)/k. As there are (k-I) degrees of freedom in each exposure, this
variance will have m(k - I) degrees of freedom. If for each one of the m exposures
a mean density is found, this will be distributed with a variance of 4+ (4+ 4 ) j k
which will have (m- I) degrees of freedom. The total number of samples is
mk = n. We can then construct the following table :
Degrees of freedom Value of variance

Between exposures m--I 2+T+ u;

Within exposures dk--I) 4+d


Total mk- I (F) <u:+ut>+.”,

Whether the two estimates of the variance differ significantly can be tested by
calculating Fisher’s parameter 2 ( =half the difference of the natural logarithms
of the two estimates) and then finding the probability of exceeding this value by
chance on the given degrees of freedom from tables given by Fisher (1946).
Table IV gives the values of the variances, the respective degrees of freedom, the
parameter x and the probability P in the case of Chalk, Silurian limestone, Old
Red Sandstone and Lower Mottled Sandstone.
The values of P clearly show that there is no difference among the several
exposures in one rock which is not accounted for as a random sampling effect of
the differences within each exposure. In other words, there is no “regional
variation of density” for the rocks in Table IV.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
266 D. S.ParasniS
TABLE
I1
Densities and Porosities of Rocks
Saturation
Specimen Apptnximate density €
No. lithology Locality per cent
Ps
g b '
I Limestone Woolhope Cockshoot 2.67 2.64 0.03 2.72 3.1
2 Keuper 8.8. 2-13 1-81 0.32 2.67 47'0
3.1 Sandy 1.s. Hanley William 2.56 2.47 0-og 2.72 9'9
3 '2 shaly S.S. Hanley William 2-56 2.47 0-og 2-72 9'9
3 '3 Marly shales Hanley William 2-53 2-41 0.12 2.74 14.0
3 '4 calcareous 8.8. Hanley William 2-62 2-59 0.03 2.67 3'1
4'1 Calcareous shales Nr. Woolhope 2'55 2-45 0.10 2.72 12'2
4'2 Shale Nr. Woolhope 2.5 I 2.43 0.08 2.64 8 '7
5'1 Calcareous shales Outside Woolhope 2.58 2-50 0.08 2.72 8.7
5 '2 Limestone Outside Woolhope 2.68 2-65 0.03 2.73 3-1
6'4 Siltstone Little Malvern 2-62 2.56 0.06 2.72 6.4
7'1 Carbonaceous 8.6. Wyre Forest 2-50 2'35 0.15 2.76 17.2
8.1 Shale Nr. Highley 2'49 2.32 0.17 2.80 20'5
8.2 Sandstone Nr. Highley 2-41 2'22 0.19 2.74 23'5
8.3 Sandstone Nr. Highley 2'43 2.27 0.16 2.70 19.0
9 Calcareous 8.8. (Keele
beds) Nr. Chelmarsh 2.46 2-31 0.15 2-72 17-2
I0 Lower Mottled 8.8. Nr. Quattford 1.26 2.01 0.~5 2-68 33'3
I1 Calcareous 8.8. (Enville
beds) Nr. Tuckhill 2'33 2.11 0.22 2.71 28.2
I2 Mottled 8.8. 2'22 1-95 0.27 2-68 37'0
13 oolitic 1.8. Nr. Upton St. Leonards 2.36 2.20 0.16 2-62 19.0
14.1 Liassic 1.s. Aust Cliff 2-56 2.46 0'10 2-74 11.1
14-2 Liassic I.s. Aust Cliff 2'55 2.44 0'11 2-74 12-4
14.3 Green Marl Aust ClifF 2.38 2-13 0.25 2.84 33'3
14'4 Green Marl Aust Cliff 2 '40 2.13 0.27 2-92 37'0
14'5 Blue Paper shales Aust Cliff I '90 1.63 0.27 1-24 37'1
14.6 I Red Marl Aust Cliff 2-44 2-26 0.18 2.76 22'0
14.62 Red Marl Aust ClifF 2-42 2-23 0.19 2.76 23 '5
14.63 Red Marl Aust Cliff 2'44 2.26 0.18 2.76 22'0
1464 Red Marl Aust Cliff 2-44 2.25 0.19 1-78 23 '5
I5 Keuper 8.5. Nr. Newent 2.26 2.04 0.22 2-62 a8'2
16.1 Metamorphic (gneiss) Malvems 2.62 2.59 0.03 2-67 3.1
16.2 Metamorphic (gneiss) Malverns 2'73 2-72 0.01 2.75 I '01
16.3 Metamorphic (gneiss) Malverns 2.64 2.60 0.04 2-71 4.2
16.6 Metamorphic (gneiss) Malverns 2.64 2-60 0.04 2-71 4'2
16.7 Metamorphic (gneiss) Malverns 2'75 2.72 0.03 2-80 3'1
17'1 O.R.S. May Hill (nr.Monmouth) 2.61 2.56 0.05 2-70 5 '27
17.2 Sandy Marl May Hill (nr.Monmouth) 2.60 2-53 0.07 2.72 7-51
18.11 Carbonaceous 1.s. Nr. Lydbrook (Forest
of Dean) 2.66 2.64 0.02 2-70 2'02
18.12 Carbonaceous 1.8. Nr. Lydbrook (Forest
of Dean) 2.71 2.69 0'02 2-74 2'02
18.2 Carbonaceous 1.s. Nr. Lydbrook (Forest
of Dean) 2.64 2-60 0.04 2.71 4-15
18.3 I Carbonaceous 1.s. Nr. Lydbrook (Forest
of Dean) 2.64 2.62 0.02 2.68 2 *02
18-32 Carbonaceous 1.s. Nr. Lydbrook (Forest
of Dean) 2.67 2.65 0.02 2-71 2'02
19 Silurian 1.s. Wenlock Edge 2.65 2.63 0.02 2-69 2
'-
20.1 Calcite Vale of Severn (nr.
Presthope) . 2.70 2-69 0.01 2.72 1-01
20-2 Limestone Vale of Severn (nr.
Presthope) 2.65 2.61 0.04 2-72 4-15
20.3 Limestone with calcite Vale of Severn (nr.
Presthope) 2 -69 2-68 0.01 2-71 I *ox

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
A study of rock densities in the English Midlands 267
TABLE 11-continued
21 Limestone West of Much Wenlc~ c k 2.61 2'73 7'5
AI Chalk Nr. Cambridge 1'93 2.50 61.4
Az Chalk Nr. Cambridge 1-97 2.54 58.8
A3 Chalk Nr. Cambridge I '94 2.62 72.5
BI Chalk Nr. Cambridge I '94 2.45 54'0
BZ Chalk Nr. Cambridge I -96 2-50 56.4
CI Chalk Nr. Cambridge I .96 2.66 72.5
c2 Chalk Nr. Cambridge 1.95 2-67 75'5
c3 Chalk Nr. Cambridge 1-95 2.61 69.5
Notes: Specimen numbers have no textual significance.
S.S. =sandstone ; 1.s. =limestone.
O.R.S.=Old Red Sandstone.
Letters A, B, C denote pits from which the Chalk samples were obtained.

TABLE
111
The Errors in Porosity
I -A =€
g/m3 per cent
1'00 0.014
0.75 0.025
0.50 0.056
0.25 0.224
0'10 I -40
0'05 5 -60
0'00 03

TABLE
IV
The Analysis of Variance
Variance
Rock formation 2 P
(S/cm")"
Between exposures z OWOOI
Chalk 6 0.0008
Between exposures 4 0.0007
silurian b a t o n : { w i ~ nexposures 5 o*ooIo
0.178 %s per cent
Between exposures I 0.0005
O.R.S.
exposures
Between exposures
3
I
o*0013
0-0008
} 1-131 >5 per cent

1.040 $5 per cent


0'0001

3.5. Discussion of sampling variances.-The specimens in Table I1 naturally


fall into the following five groups : (I) sandstones with four sub-groups, (2)lime-
stones with three sub-groups, (3) marls with two sub-groups, (4)shales, and
(5) metamorphic rocks. Table V gives the mean saturated density of each group,
the observed standard deviation and the number of degrees of freedom on which the
variance has been calculated. The corresponding quantities for the dry density
could also be given but it is more useful to consider the porosity, and the table also
contains the necessary entries in this respect.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
268 D. S . Parasnis
TABLE V
Sampling Variance of Density and Porosity
Standard Degrees Standard Degrees
of
No. Lithological Mzan deviation of Mean deviation
freedom
OE
V

I Sandstones
Ia Old Red 2.58 0.035 5 7" 2.46 4
Ib Carboniferous 2'42 0.061 5 21.95 4.84 5
IC New Red (all) 2'22 0.061 3 36.4 7'95 3
Id Lower Mottled 2'247 0.023 2 ... ... ...
2 Limestones
za Silurian 2 -65 0.034 8 4'31 I -88 9
2b Carboniferous 2.66 0.029 4 2'44 0.95 4
2C Chalk 1 '95 0.0126 8 65.13 8-57 8

3 MlUlS 2-42 0.025 5 ... ... ...


3a Red Mark (Aust Cliff) ... ... ... 22-75 0.87 3
3b Green (Aust Cliff) ... ... ... 35'15 2'57 I

4 Shales (all) 2.56 0.055 6 ... ... ...


5 Metamorphic rocks
from Malvem Hills 2.69 0.056 4 3'55 I -96 4

Several interesting conclusions can be drawn from this analysis. In the first
place, Silurian and Carboniferous limestones seem to form a remarkably uniform
deposit despite their different ages. Sandstones, on the other hand, are more
variable, but sandstones of the same age group show a small variance of density.
The porosity of sandstones even of the same age group has a rather large variance.
In the case of New Red Sandstone, this is not unexpected judging from the condi-
tions of its deposition. Chalk appears to be the most uniform deposit so far as
density is concerned and also the most variable in porosity. As chalk is supposed
to be of shallow-water origin it might be argued that a full differentiation of the
sediments could not have taken place and consequently chalk would be very far
from being a homogeneous deposit and then the first fact would seem very sur-
prising. However, a glance at Table I1 will show that the grain density of chalk is
very variable. The outstanding fact is its low value, although pure calcite has
a density of 2.70 g/cm*. It was suggested by Dr Cook that as chalk grains may be
shells of animals, they might have cavities so minute that complete drying of a
specimen cannot be achieved. Therefore, a sample of chalk was pulverized and
dried when its grain density was found to be 2.64g/cm8 (obtained by a pycno-
meter), which seems to support the above suggestion. This value is still too low
and the difference is probably due to a number of shells not being crushed at all or
due to an adsorbed film of water on each grain which cannot be driven out except
at about 300 deg. C. or beyond. The second reason seems to be the more likely
one from a number of similar experiments made by the author on granular
materials. As chalk contains a considerable amount of soluble material the
porosity would depend very much on the water table, the dip of the strata, etc. and
would show the large variance that has been found. The variances of density and
porosity of the metamorphic rocks are rather surprisingly small. But the low

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
A study of rock densities in the English Midlands 269
values of porosity are not unexpected from the nature of the rocks. The slightly
higher grain density of green marls may be due to the presence of glauconite. The
difference of density between Silurian limestone and Carboniferous limestone
(za, 2b in Table V) is not significant, but that between the Old Red Sandstone and
the Carboniferous Sandstone is.
4. Comparison of theJield and laboratory methods
4.1. In order to test whether the density values found by the two methods refer
to the same population of rock formations in each case, Table VI was constructed.
In this table the following details are given : the locality, the type of rock, the values
of density obtained by the field method (H),its standard deviation, the laboratory
measurement (L), its standard deviation, the difference between the two density
values, the standard deviation of the difference (a),and the ratio (H-L)2/a2.
is calculated and the probability is found of x2 exceeding a
x2 ( = X [ ( H - L)2/a2])
certain value on eight degrees of freedom (the actual number) from Fisher's tables
(Fisher, 1946). It is found that the probability is 0.2, which means that H and L
are not significantly different from one another. No correlation can be found
between (H- L)2/02and either lithology or porosity.

TABLE
VI
Comparison of the Two Methodc
Densities in g / m *
Field Std. Lab. Std. H--L
Locality Rock value dev. value dev. 0 (ff-L)'/d e(permt)
type H H L L
Woolhope
Cockshoot Silurian 1.s. 2.73 0.25 2.67 0.034 +0.06 0.25 0.058
Hanley William O.R.S. 2.47 0.048 2.58 0.035 -0.11 0.06 3.38
May Hill (Mon-
mouth) O.R.S. 2.46 0.08 2.58 0'035 -0.12 0.09 1.78
Wyre Forest Coal Measures 2.70 0.10 2.42 0.061 +
0.28 0'12 4-71
Forcat of Dean Carb. 1.8. 2-54 0'073 2.66 0.029 -0.12 0.14 0.74
Bridgnorth ClSs Triassic 9.9. 2-30 0.13 2.25 0.023 +
0.05 0.15 0.11
Upton St. Leonards Lias 2.28 0.19 2-40 0.05 -0.12 0.20 0.36
Gog Magog Hills
(Cambridge) Chalk l*92* 0'20 1-95 0'012 -0.03 0'20 0'02

Degreea of freedom 8 x*= I I *16


* From measurementa made by Cook and Thirlaway (unpublished).
4.2. Measurements in a mine.-The value of the density of a formation can
also be found by measuring the difference of gravity at two points in a mine shaft.
Such measurements have been made by Cook and Thirlaway (1951) and Table VII
gives the analysis of these. They are grouped separately because they are not
directly comparable with the hill measurements. Mine measurements are
expected to be more accurate because there is no error arising from a horizontal
gradient of gravity, and a comparison of Tables VI and VII shows that this is so.
The conclusion is that M and L do not differ significantly.
5. Conclusions and catalogue of densities
The study of sampling variances described in this paper shows that in gravity
surveys in the Western Midlands, laboratory measurements on rock densities will
be sufficiently accurate even if hill measurements are dispensed with. This

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
270 D. S. Paramis
conclusion is completely different from that reached by Hammer (1950)from
density measurements in America. However, his results comprise (repeated)
gravimeter measurements in a mine shaft and determinations on several samples
from the shaft and a borehole in the vicinity and do not include any density profile
determinations.
TABLE VII
Density Measurements in Mine Shafts
(Observers : Cook and Thirlaway)
Field Standard Laboratory
Locality of mine Formation value deviation value uL M-L u -
(M-L)'
'
U
*M L
S. Wales and Forest of
Dean (Mean) Pennant grit 2.65 0.05 2.67 0.01 -0-02 0.05 0.16
S. Wales (Britannia Pit,
Pengam) Deri Beds 2.77 0'02 2.73 0.06 +0.04 0.064 0.39
Alveley Pit Coal Measures 2-52 0.01 2-42 0.06 +O-IO 0.06 2.78
Degrees of freedom 3. x8=3*33
P (from tables) >0-3.
In the sampling of rocks, attention must be paid to lithology as variances tend
to vary with it. Table IV is based on the values of density when the specimens are
saturated with water and the results show that this is the most likely condition in
nature. As no regional variation in density was found, a sample collection from
one exposure should be sufficient for laboratory measurements. It must be
pointed out that this conclusion is only true for the particular district studied in
this paper, and it is not proved that there may not be considerable variation in
density in the same formation in other districts. Table VIII has been formed for
general use in gravity surveys of the West Midlands. The adopted values are the
mean of the field value and the saturated density in laboratory measurements.
Chalk has also been included from the point of view of the completeness of the
paper.
TABLE
VIII
Catalogue of Densities
Adopted
No. Formation density
(f3/=*)
Siluriw limestones and shales 2-70
Old Red Sandstone 2'52
Carboniferous limestone 2 -60
Middle Coal Measures
Shales and Sandstones 2.48
Keuper Sandstone 2.27
Keuper Marl 2.42
Upper Lias 2-34
Chalk 1 '94
Malvernian Gneiss 2.69

Notes on variability :
(I) Silurian rocks and Carboniferous limestone do not show wide variation in
sampling.
(2) Sandstones show a wider variation than limestones.
(3) Chalk is very uniform in density.
(4) The density of metamorphic rocks is applicable to the Malvern area only.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018
A study of rock densities in the English Midlands 271
6. Acknowledgments
The author wishes to thank Dr A. H. Cook for his advice and help in the
accumulation and interpretation of the data. He is grateful to him and to
Mr H. I. S. Thirlaway for making their mine measurements available for this
paper. The author also wishes to thank the Royal Dutch-Shell group of Oil
Companies for a studentship.
Department of Geodesy and Geophysics,
Cambri4ge University:
1951June 12.

References
Cook,A. H. (195I), “An investigationof the errors of some recent gravimeter observations”,
M.N., Geophys. Suppl., 6, 194-208.
Cook, A. H. and Thirlaway, H. I. S., (1951)“ A gravity survey of the Bristol Coalfield area”,
Q. J. Geol. SOC. (in press).
Fisher, R. A. (1946), Statistical methods for research workers (London).
Hammer, S.(1939),“ Terrain corrections for gravimeter stations ”, Geophysics, I, 184-194.
Hammer, S. (1950), “ Density determinations by underground gravity measurements ”,
Geophysics, 15,637452.
Holm-, A. (1921),Petrographic methods (London).
Kidston, R., Cantrill, T. C. and Dixon, E. E. L. (1917),“The Forest of Wyre and Tither-
stone Clee Hill Coalfields ”, Trans. Roy. SOC. Edin., 51, 999-1084.
Moore, C. (1902),“ The study of the volume compositions of rocks ”, Proc. Geol. SOC.
Liverpool, 9, 129.
Nettleton, L. L. (1939),‘‘ Determination of density for reduction of gravimeter obser-
vations ”, Geophysics, I, 176-183.
Wickham King, W. (1925),“ Notes on the Old Red Sandstone of Shropshire ”, Proc. Geol.
Assoc., 36, 383-389.

Downloaded from https://academic.oup.com/gsmnras/article-abstract/6/5/252/561095


by Kyushu University user
on 30 May 2018

You might also like