Earth Sciences / 75irregu-lar process, so a thick layer may have accumulated rapidly, a thin one more slowly, and
there is no way to tell. Some sediments, however, build more regularly, and it was the record
they left that allowed the retreat of Scandinavian ice sheets, starting about 10000 years ago, to
be traced. Each spring, as the ice melts, an assortment of mineral particles is washed into a lake
by the meltwater. Heavier particles, such as sand grains, settle quickly. Later in the year, as
water freezes again, the supply to the lake ceases and the finer particles, of silt and clay, gradually
settle on top of the sandy layer. Year after year the process is repeated, each pair of layers, one
pale and coarse, one dark and fine, being known as a ‘varve’. These can be counted, each varve
representing one year, and if varves are forming at the edge of a retreating glacier they will
follow it, so that its progress can be traced and dated. The study of varves is known as varve
analysis, varve chronology, or a varve count.
Varves resemble tree rings, which provide another method of measuring time. In spring, woody
plants grow rapidly by producing large, thin-walled cells in the xylem, just below the bark of stems
and branches. Growth slows in summer, ceasing in late summer, and consists of smaller cells with
thicker walls. The large cells of spring are pale in colour, the smaller ones of summer dark, and so
each year the plant produces a ring of pale wood separated by a thin, dark ring from the pale wood of
the following year. A count of the rings is a count of years, but there are some risks. If conditions are
very severe, a plant may produce no growth for a whole year, and if conditions are unusually favourable
it may produce two or more sets of rings. For this reason, tree-ring dating (called dendrochronology)
must be based on as many specimens as is practical, obtained from widely scattered locations. The
fact that rings are strongly affected by growing conditions has advantages. The width of rings can be
used to infer weather (dendroclimatology) and environmental (dendroecology) conditions at the
time they formed.
Obviously, the study of tree rings can provide dates only up to the age of the living plant from
which they are taken, but trees can live a surprisingly long time. There are bristlecone pines
(Pinus longaeva), found in California, more than 4600 years old, and correlating rings from them
(taken as cores, without destroying the tree) with rings from dead pines has allowed scientists to
construct a chronology for arid zones going back 8600 years and, at the upper tree limit on
mountains, one going back 5500 years.
These chronologies are used to calibrate radiocarbon (^14 C) dates. Bombardment by cosmic radiation
generates neutrons, a few of which collide with atoms of nitrogen (^14 N), displacing a proton and
converting the^14 N to^14 C. Chemically,^14 C behaves just like ordinary^12 C and living organisms exchange
both with their surroundings. When they die, however, carbon exchange ceases. Carbon-14 is
radioactive, half of any amount of it decaying to^12 C in 5730±30 years (its half-life), so the ratio of
(^12) C: (^14) C in dead organic matter is directly related to the time that has elapsed since it died. Radiocarbon
dating rests, however, on the assumption that the rate of^14 C formation in the atmosphere is constant.
This is now known not to be so, because the intensity of cosmic-ray bombardment is variable, but,
when correlated with tree-ring series from bristlecone pine, radiocarbon analysis makes it possible
to date material up to about 70000 years old.
Dating material older than this requires other methods. These, too, are based on the decay of radioactive
elements, but ones with much longer half-lives. The first to be exploited were uranium (U) and
thorium (Th). Uranium occurs naturally as a mixture of two isotopes,^238 U and^235 U in the constant
proportions 137.7:1; both decay to stable isotopes of lead (Pb). Uranium-238, with a half-life of
4510 million years, decays to^206 Pb, and^235 U, with a half-life of 713 million years, to^207 Pb. Thorium-
232, with a half-life of 13900 million years, decays to^208 Pb. Lead also occurs naturally as the stable
isotope^204 Pb, so this must be deducted from lead isotopes resulting from radioactive decay before an
age can be calculated.