Page 21 - Petrology of Sedimentary Rocks

COLLECTION AND PREPARATION OF SAMPLES FOR ANALYSIS

Sampling

There are two methods of sampling, the channel sample and the spot sample. The
channel sample is useful when you are trying to get average properties over a large
stratigraphic interval. For example, if you are making a channel sample of a sand
exposed in a 20-foot cliff, you can dig a trench the full height of the cliff and take a
continuous sample all the way up, mix it up thoroughly and analyze the whole thing
together, or you can take small samples every foot, mix them up thoroughly and analyze
the mixture. This method is good for determining economic value, for example if you
would like to find out how much iron or how much clay mineral is present in the whole
cliff. But it is absolutely useless for determining origin or sedimentational conditions,
since you are analyzing a mixture of many layers deposited under varying conditions.
We will therefore use spot sampling. To take a spot sample, one selects a representa-
tive small area and takes a sample of just one bed that appears homogeneous--i.e., you
sample just one “sedimentation unit.” For example, if an outcrop consists of two-inch
beds of coarse sand alternating with one-inch beds of fine sand, your spot sample will
include only one of these beds, trying to get it pure without any admixture of the
alternate bed. The smaller your sample is, the more nearly will it represent
sedimentational conditions; for sands, 50-100 grams is probably adequate (this is a
volume of 2 to 4 cubic inches); for gravels, a larger sample is required to get a
representative amount of pebbles, thus usually 1000 grams or more is needed (50 to 100
cubic inches); for silts and clays, you may sample as small as 20 to 50 grams. Just
remember to sample only one layer of sediment if possible, excluding any coarser or
finer layers that lie adjacent.

If you have the problem of sampling a sedimentary body, be it a recent sand bar or
a Cambrian shale, first decide on the extent of the unit to be sampled (which may be
either area, linear distance along an outcrop, or stratigraphic thickness) and then on the
number of samples you want to take. Dividing these two numbers gives you the spacing
of the samples. If you are going to take only a few samples, it is a general rule to take
representative ones (i.e., sample the most common or typical beds exposed). If you are
going to take a large number of samples, the ideal way to approach it is as follows: (I)
after dividing the extent of the unit by the number of samples, set up an equispaced
grid of sampling stations through the unit. For example if the cliff is 20 feet high and
you want to take about 7 samples, then 20/7 = 3 and you space your main sample
locations at 1.5, 4.5, 7.5, 10.5, 13.5, 16.5, 19.5 feet and sample the material at those
precise points. By doing this according to a rigid grid spacing, you are pretty sure to get
representative samples of the typical rock. In addition to the samples spaced on the
grid, you should take extra samples of different or peculiar layers that may have been
missed on the grid; in the example above, if a conglomerate layer had occurred at I2
feet you would take an extra sample of that even if you did not land on the grid, and
you’d end up with 8 samples of the outcrop. These extra samples should be especially
labeled inasmuch as they are somewhat unique and not representative of the outcrop.
The whole purpose of the grid sampling method is to avoid the error of non-
representative samples; if one were to sample a batholith, for example, one would not
collect 40 samples of pegmatites and only 5 samples of legi timate granite no matter
how monotonous and similar the granite appeared; one would collect chiefly granite
samples with only a few pegmatites. Above all, use your head and in the samples you

I5