Stating seed longevity is inherently problematic. Seed longevity is usually reported in literature sources as the percentage of seeds alive after some given number of years or just as the existence of viable seeds after some number of years. Mostly we have echoed this in the entries in the database. However, such numbers do not reflect the nature of seed mortality in the soil. Unlike healthy humans that mostly die of old age, seeds in the soil tend to die at a constant rate. One way to understand that is this: if weed seeds had life insurance policies, their payments would be the same regardless of the age at which they bought the policy. Consequently, seed mortality is best expressed as either the percentage of seeds that die in a year, or the half-life of the seed population. Where data are available to express mortality in this way, percentage loss per year is given. Seed survival is highly dependent on weather (particularly for recently shed seeds), the presence or absence of seed predators that like the particular species, and soil and management conditions, so all types of data should be used only as a general indicator of the species persistence in the soil.
Weed seeds die primarily in one of three ways: (i) they begin to germinate in conditions that do not allow establishment, (ii) they are eaten by seed predators, or (iii) they die from physiological breakdown. The relative importance of the three mechanisms is generally believed to be in the order listed, though most of the species that have been studied systematically have been grasses with relatively large seeds.
Although most weed species possess mechanisms for determining appropriate seasons and conditions for germination (Timing of germination, Tillage & germination), these mechanisms do not work perfectly. Consequently, many seeds germinate too deeply in the soil for successful emergence, they begin to germinate and then dry out and die, or they begin to germinate and are then attacked by soil organisms.
In general, seed predators are most effective against seeds that are on or near the soil surface. Consequently, if weeds have gone to seed in the garden, fall tillage will tend to protect the weed seeds. Ground foraging birds and mice tend to be the major predators on seeds greater than 2 to 5 mg, whereas major predators on smaller seeds include ground beetles of the carabid family. Earthworms consume grass seeds and actually digest a substantial proportion of them; their action on broadleaf weeds has not been studied.
Although seeds are inanimate creatures, they are alive and metabolically active. In the soil, most weed species persist in a moist condition and are capable of metabolic repairs. Eventually, however, damage to membranes, genetic mistakes, toxins and other metabolic problems cause loss of vigor and eventual death. This happens most quickly at warm temperatures and when the seeds are nearly but not completely dry. Since both of these conditions are most likely to occur near the soil surface, seeds near the surface that remain dormant are in a risky position.