X-Ray Analysis of the Gingin Upper Greensand. 
233 
For other experimental details regarding mounting of powders, films, etc., 
reference may be made to a previous publication (4). 
ACC'URAC'Y OF RESULTS. 
In tliis work the determination of d /n by a graphical method pi’eviously' 
described (4) was not resorted to. 1’he intei-planar spacings v ere determined by 
a direct calculation from the Itragg equation nX - 2d sin 0. This was rontit'red 
possible because quartz present in the coai'se fraction of each sample produced 
characteristic blotchy lines whicli could easily be recognised and used for cali- 
bration. I’ho lilotehy characti'r of the lines was a.s.sociatefl with the jjarf icle 
size. It was found that r (the radius of curvatiire of the film) was constant 
along a film but varied from film to lihn. This variation was small and its 
extreme range was not moi-e than 0 • 5 per cent. 
Accuracy of Results for the Coarse Fractioyis. — -Values of r calculated on the 
basis of the three t[uartz lines of interplanar spacing 3*34, U814 and 1*538 A 
were found not to vary by more than 0*25 per cent. Such a. variation lies 
within the limits imposed by inability to measure any 2,5 (the separation of 
corresponding diffi*action lines along the film in the etpiatorial plane) to any 
greater accuracy than 0* 1 mm. Hue then to an error in 2s of 0* 1 mm. in lines 
used for calibration an error of 0 • 25 per cent, aiul 0 • 09 per cent, is possible for 
any value of tl n calculated ovta- tlie range 10 A to 1 A. 
On the assumption that there is no error in r involved in calibration 
measurements and that r is uniform across the film, the error in d /n associated 
with an error of 0* 1 mm. in 2s over the range from 10 A to 1 A varies from 
l’]7 per cent, to 0*075 per cent. 
Hence the total maximum percentage error, assuming (a) uniform radius 
along a film and (b) d(2.9) 0* 1 jnm. in calibration and in measurement of an 
unknown line, is 1*42 p<n’ cent, to 0- 10 per ctnt. over the range 10 A to 1 A. 
\^'eak lines may have a great(T ptTcentage error. 
On this basis the maximum possible error in d /n at 2*5 A becomes 0*01 A 
so that for smaller spacings than this, d/n is ro^'orded to O-OOl A. 
Accywacy of Results for the Fine Fractions,- In the tilms of the tine fractions 
the absenf*e of blotchy lines rendered the certain recognition of (jiiartz impos- 
sible. ConseejLiently, as no lines wt're available for calibration purposes, it 
was necessary to assume a constant radius of curvature for all films. The 
figure ado])ted was 2*798 cms. 
In the abse'Hce of glauconite a line appearing with interplanar s])acing 
3*34 A is eonsider't'd due to (juartz. From an examination of Table I. it is 
seen that the maximiun \-ai*iation in d n for this line in the patterns of the 
tine fractions is (>*(H A. Tliis is due to variations in 2s of 0*(d Tum., corres- 
ponding to the maximum en'or (‘onsicU'n'd possible in measurements of 2s for 
all but the \ ery I’aintest liiu's. Hence, on this basis, it is considered that the 
assumyition of a constant radius of 2.798 cms. yields results for the finer frac- 
tions wliich are no less inaccurate +han those for the coars(T fractions. 
X-RAY DATA. 
In Tabh^ I. are tabulated all obs(‘r\'e<l values of d 'n for the samples exam- 
ined. This inchuk's the patterns obtained for samples which were heated in 
an electri(‘ furnace in air to 500®C\ for 48 hours. 
It has been ]ireviously pointed out (4) and (9) that the commonly occurring 
minerals in soil colloids inclu(l(‘ the clay minerals, cuartz, and the oxides and 
hydroxides of A1 and of F(\ 
