A sort of "annual growth ring" in the sediment of a lake bottom.
The tremendous impact of a glacier upon the landscape is not limited to its action while in a solid form -- when the ice melts, large amounts of water need to have somewhere to go. As you might expect, water often collects into lakes in the glacier's proximity.
As the seasons pass, the glacier will melt to a certain extent during the summer, and less (or not at all) in the winter. Whatever the amount, water is carried into a glacial lake1 at a much greater rate during the summer than the winter.
All runoff water carries sediment of some kind. As the water enters the lake, the sediment it carries is deposited on the bottom. The greater the volume of water, the more it is able to carry, and the larger the particle size. Thus, the spring and summer meltwater carries a much larger volume of material into the lake, and a large amount of relatively coarse sediment is deposited on the bottom. During the winter, the lake bottom
receives a much smaller amount of sediment, which is much finer. Sometimes this winter sediment comes from the lake itself -- the lake's water is much more still (perhaps it freezes over), allowing organic particles and very fine clay particles to settle to the bottom. The winter sediment's constituents usually make it much darker than the summer sediment it covers.
Over the years, alternating dark and light layers accumulate on the bottom of the lake. These are called "varves" or "varved sediments". The former bottom of glacial Lake Agassiz left an immense deposit of varved sediments, covering most of southern Manitoba and western Ontario, with extensions into Saskatchewan, Minnesota, and North Dakota.
Some lakes have thousands of these layers at the bottom, allowing them to be used for dating events in much the same way as tree rings. Although absolute dating of this kind is useful, incomplete sequences of varves can be fit in to the overall picture using fossil organic material (usually pollen) in the dark winter layer. "Varve analysis" as a form of geochronology was pioneered by the Swedish geologist Gerhard Jakob De Geer in the late 19th Century.
1This isn't really limited to glacial lakes, but any body of water where spring or summer meltwater (from glaciers or snow) plays a major role.