One of my faculty colleagues, Michael Runtz, took this photo (left) of ice bubbles in Cranberry Lake in Ontario. How did the bubbles form in this amazing fashion?James Cheetham, Department of Biology, Carleton University, Ottawa, Ontario, Canada
Bubbles rise constantly from the sediment on the bottom of the lake. The ice that forms does so layer by layer (night after night, cause then it's coldest) and the (probably methane) bubbles that rise get trapped in the newly formed layer of ice. If you like to find out if it is methane that is trapped inside the bubbles: make a tiny hole in the ice directly to the trapped gas inside the bubble and light it. If it's methane you'll see a flame...
You don't need to go to Canada - I've photographed similar frozen bubbles in puddles in this country:http://www.flickr.com/photos/limegreensport/2270561041/
A body of water freezes from the top downwards. Bubbles of gas rise from the bottom. So, what you're seeing frozen in the ice is bubbles of gas that have "hit" the underneath of the frozen layer; flattened against it; and then that layer of water has frozen around the bubble. Then, the next bubble of gas has risen from the bottom and hit the bottom of the newly frozen layer and so on and so on...
The following answer has been selected and edited by New Scientist staffWithout any scale reference or indication of the depth of the lake I cannot tell for certain, but I have seen similar bubbles frozen in ponds where I grew up in upstate New York.In freshwater ponds and lakes, the biological activity of microbes in the sediments on the lake floor produces bubbles of gas, usually methane or carbon dioxide. In winter this activity is slow, but it is still present.The gas bubbles rise to the frozen surface of the lake, becoming trapped there. The following night, another layer of ice forms beneath the bubble, so it is encased in ice. This leads to the flattened shape you see. The picture is a frozen daily record of the gas emissions.Obviously you need calm, shallow-water ponds or the sheltered edges of small lakes in order for this phenomenon to occur.James C. Field, Aberystwyth, Ceredigion, UK