Recently, in my historical geology class (historical geology is the history of the Earth, by the way, not the history of geology) I discussed a famous rock unit known as the Gowganda Formation. The Gowganda Formation consists primarily of conglomerates (diamictites, really, but most people aren't familiar with that term) and laminated mudstones. Here's a picture of one of the conglomerates I picked up once on a field trip in Michigan (it's been glacially polished and rounded by transport during the last ice age).
This picture really doesn't do it justice - it's a beautiful rock. The conglomerate is not just any conglomerate either, it represents special type of rock called a tillite. When large ice sheets from glaciers move across the landscape they erode the bedrock and pick up material of all sizes from microscopic clays to large boulders. When the climate warms, and glaciers melt, they drop all of this unsorted geologic material into deposits known as glacial till or simply till. Till, over periods of geologic time, can be lithified into a sedimentary rock known as tillite. Tillites are therefore representatives of past ice ages.
Here's a nice picture of the Gowganda tillite in the field.
The other major rock unit in the Gowganda Formation are laminated, or varved, mudstones. These also form in environments characterized by glaciation. As glaciers are melting, large ice margin lakes form which may be open in the summer and frozen in the winter. This results in differences in winter (finer clays) and summer (coarser silts and sands) sedimentation on the lake bottom leading to yearly couplets of lighter and darker laminations called varves. Counting the varves allows one to determine how long the ice-margin lake existed.
The outcrop above, in a quarry just north of Lake Huron and east of Sault Ste Marie (46.413° N, 83.325° W) in Ontario, shows at least 750 years of varve couplets in a mudstone (also called an argillite by geologists). Also shown, with the black arrow in the image, is a granite dropstone - the same granite seen in the tillite above). Here's a better picture of a dropstone in the Gowganda varved mudstone sequence.
Dropstones form when chunks of ice break off from the glacier, float into the large ice-margin lake, and melt. When they melt, they'll drop anything their carrying including, at times, sizable rocks which become dropstones in the lakebed sediments as seen above.
The two lithologies (rock types) of tillites and argillites, along with the varves and dropstones in the argillite, clearly indicate glaciation at the time these rocks were forming. Some locales even show multiple advances of ice (tillites separated by thick argillite sequences). So when did this glacial event occur?
Plate tectonics moves continents around over time so another reasonable question would be to ask where this area was during this period of time. Was it near the poles (in which case glaciation would not necessarily be surprising) or was it at lower latitudes (which would indicate a global ice age during the time)? Evidence from paleomagnetism (studies of the ancient magnetic field in rocks indicating their magnetic latitude at time of formation) seems to indicate the Gowganda formed at relatively low latitudes. In other words, at the time of a glacial ice age.
A digression... Most people know about THE ice age. That at some point in the past (most people you ask on the street would say millions of years ago but the last great ice sheet melted from the Hudson Valley only about 12,000 years ago), the Earth was covered with massive ice sheets. What most people don't know, however, is that the Earth has had multiple ice ages in its 4.5 billion-year history. The one we think of as THE ice age is only the most recent which occurred in a period of time known as the Pleistocene Epoch (beginning only 2 million years ago).
The earliest glacial ice age we know about is the one that formed the sediments of the Gowganda Formation in Ontario, Canada. Similar age glacial sediments exist in western Austrlia (Hammersley Province) and southern African (Transvaal Basin) and elsewhere indicating a global event (even though those places weren't in the same geographic position they are today because of plate tectonic movements). This glacial event has been called the Huronian Glaciation (because the Gowganda Formation is found on the north side of Lake Huron).
So why did it take a couple of billion years of Earth history until we had our first glacial ice age? Recent thinking invokes two probably causes. The first is the decrease in methane (CH4) in the Earth's atmosphere by increasing oxygenation by photosynthetic cyanobacteria (the earliest Earth had no free oxygen in the atmosphere - it's not until the development of life that the Earth's atmosphere began to significantly change). The second cause may be due to drawdown of carbon dioxide (CO2) out of the atmosphere by the weathering of silicate minerals due to increased rifting of continents and extrusion of extensive lava flows (flood basalts) at the time. Both methane and carbon dioxide are greenhouse gases and reducing their abundance in the atmosphere would have cooled the Earth's climate significantly.
We have abundant geological evidence throughout Earth's history that changing the Earth's atmospheric chemistry will change the Earth's climate - sometimes significantly. Something to think about as we pump gigatons of CO2 per year into the atmosphere by burning fossil fuels.