How Do We Know What Grew Here? Counting Grains in Mud

The outer wall of a pollen grain is built from an extraordinarily resistant material that withstands acids, decay and the passage of very long periods, which is why grains persist in sediments where everything else organic has broken down. Plants produce it in enormous quantities, particularly those relying on wind for pollination, and it disperses widely before settling, so a lake or bog collects a sample of what was growing in the surrounding region rather than only at the spot itself. Crucially, the wall is sculptured in patterns characteristic of each plant group, so a grain can be identified under a microscope to family and frequently to genus, which turns an accumulation of dust into a readable record of vegetation.

Extracting a record from a site follows a standard procedure:

• •A core is taken from a lake bed or peat bog, preserving the layers in order
• •Samples are taken at intervals down the core and treated chemically to remove everything but the grain walls
• •Grains are counted and identified under a microscope, typically several hundred per sample
• •Results are plotted as percentages of each type against depth, giving a diagram
• •Radiocarbon dating of material at known depths converts depth into age
• •Charcoal fragments are counted alongside, giving a record of fire

The resulting sequences record changes that are dramatic when read at the right scale. Cores from northern Europe and North America show tundra species giving way to birch and pine and then to mixed deciduous forest as the last glaciation ended, with each tree type appearing at a different time as it spread north from its refuges, which is how the rates and routes of postglacial recolonisation were established. Sudden declines in particular species mark events, including a widespread and abrupt drop in elm pollen across northwest Europe around six thousand years ago that has been attributed variously to disease, to climate and to human clearance. The appearance of cereal pollen and of weeds associated with disturbed ground marks the arrival of agriculture, frequently accompanied by a drop in tree pollen and a rise in charcoal.

Pollen is one of several proxies read from the same sediment and combining them is standard practice. Charcoal fragments record fire, with particle size indicating whether the fire was local or distant. Diatoms, single-celled algae with silica skeletons, record water chemistry and depth. Chironomid remains, the head capsules of midge larvae, are strongly temperature dependent and give a quantitative estimate of summer temperature. Spores of a fungus that grows on dung indicate the presence of grazing animals, which distinguishes a landscape opened by livestock from one opened by fire. Isotopes in the sediment itself carry further information. Reading several proxies from one core allows conclusions that no single one supports, and disagreements between them are frequently the most informative part of the result.

The limitations are important and are well understood by practitioners. Different plants produce wildly different quantities of pollen, with wind-pollinated trees producing enormous amounts and insect-pollinated plants very little, so percentages in a diagram do not correspond to the abundance of plants on the ground and must be interpreted with correction factors that are themselves estimates. Some groups cannot be distinguished beyond family level, which blurs important distinctions including between cereals and wild grasses. Pollen travels, so a grain may come from far away, and the catchment of a small hollow differs from that of a large lake. Preservation varies, and acidic waterlogged conditions are needed. And absence is weak evidence, since a plant present but poorly represented may simply not appear.

The outer wall of a grain resists decay for tens of thousands of years and is sculptured in patterns identifying the plant group, so sediment layers hold a readable record. Cores show tundra giving way to birch, pine and deciduous forest after the last glaciation, with each tree arriving at a different time. Species produce vastly different quantities, so percentages do not match abundance on the ground.