The number of seeds produced per plant also varies enormously. Relatively few seeds are produced by a coconut palm, but it has been estimated that one plant of Amaranthus graecizans, an annual tumbleweed, may produce as many as 6 million seeds.

Seedcoats are important in the longevity of seed. Seedcoats of most long-lived seeds have on or near the outside a palisade, or a Malpighian, layer made up of heavy-walled, tightly packed, radially placed, columnar cells. The cells are hard and horny. Usually they are lignified or cutinized. There are no intercellular spaces.

The palisade layer is mechanically protective and highly impervious to water and to respiratory gases. Morphologically, it is the most important structure in seed longevity; most seeds with exceptional longevity have a well developed palisade layer. This layer is not so all-important to buried seeds or to soil-stored seeds, because the soil habitat apparently provides some of the same or similar protective conditions necessary to longevity.

Little need be said here about seed extraction, cleaning, and storage, except to point out that the least damaged seeds mechanically and biologically and the best stored seeds will have the greatest longevity.

Many seeds with marked longevity look much alike. Typically they are larger and heavier than the average. Coats are thick and hard and often have a smooth, polished surface. The seeds commonly will not plump if soaked in cold water.

Some examples of well-known native macrobiotic seeds are: Gymnocladus dioicus (Kentucky coffeetree), about 240 seeds to the pound; Gleditsia triacanthos (honeylocust), about 2,840 seeds to the pound; Robinia pseudoacacia (black locust), about 25 thousand seeds; Ceanothus cuneatus (buckbrush), about 55 thousand seeds; and Lotus americanus (deervetch), about 110 thousand seeds to the pound.

Seeds of cultivated plants long used in their entirety as food commonly have thinner and weaker coats than closely related wild plants. The food seeds generally are shorter lived.

The chemical composition of many crop seeds, but few others, is well known. Seeds are sometimes classified broadly according to the kind of food reserves they store, as starchy seeds (such as those of the Gramineae, the grass family); proteinaceous seeds (for example, those of the Leguminosae, the legumes); and oily seeds (from most tree nuts and many other plants).

The classification is arbitrary, because the food reserves of seeds commonly are mixtures of various carbohydrates, proteins, and fats. I know of no comprehensive English summary of the chemical characteristics of seeds.

The longevity of the oily seeds of sugar pine (Pinus lambertiana) is related closely to the kind and amount of unsaturated fatty acids in the seeds. Another suggestion is that rancidity of seed fats in pine and other seeds varies inversely with seed viability.

The degeneration of proteins in food grains roughly parallels reduction of seed viability.

Some biochemical aspects of seed viability are known, but the precise reasons for loss of viability death of the seed are not yet clear.

The amazing thing about seeds is not that they degenerate with time but that some deteriorate so slowly.

One theory for the degeneration of long-lived seeds suggests that the various proteins slowly coagulate and denature with time and eventually cannot function in germination.

A related theory perhaps the most plausible one in the present state of our knowledge is that loss of viability is due to gradual degeneration, in the nuclei of cells, of the chromatin the material basis of heredity and of the delicate mechanism for mitosis, the process by which cells divide and increase in number.

Experiments that support this second theory show that aging, heating, and X-ray treatment of dry seeds all cause similar degeneration increased mitotic aberrations and chromosome changes and increased plant mutations and abnormalities.

Degrees of aging and of treatment cause more or less proportional increases in extent of abnormalities and mutations until all viability is lost. An extension of this theory is that the mutations resulting from aged seeds is one means by which Nature produces varying races and strains of plants and advances evolution.

Seed longevity in a broad sense is an ecologic characteristic of a plant as well as a morphologic and a biochemical one. Over the great reaches of geologic time, the biology of most plant species and their seeds has come to fit approximately the habitat in which they are characteristically found.

Some plants are primary pioneers. They grow most commonly on ecologically tough sites where soil is scarce or poor.

Other plants are secondary pioneers. They are found commonly in abundance on well-developed soil profiles from which much or all of the previous plant cover has been removed by fire, logging, or clearing. And of course there are other comparable ecologic types of plants.

What about the life history of those secondary pioneers that are killed outright by fire but produce heavy seeds not scattered by wind?

How are these plants able to revegetate a burned area promptly and abundantly, as they so often do? Simply because they have mechanically durable, heat-resistant, and long-lived seeds.

In studies of soil-stored seed under ecologically mature forests, it is common to find more seeds of the pioneer vegetation, largely displaced by ecologic development, than of the climax tree species. There is an obvious and essential ecologic value to pioneer plants from long-lived seeds that is, from seeds well distributed in time.

A few plants produce two kinds of seeds according to season and physiologic status of the plant. An example is Halogeton glomeratus, a serious weed of wildland desert ranges. One kind of seed will germinate immediately upon maturity. The second kind is dormant and will not germinate until some time after maturity. This again is distribution of seeds in time, with seed longevity as one requirement for success of the process.

Plants (particularly woody plants) characteristic of and climates are believed in general to have longer lived seeds than plants of tropical or of humid-temperate habitats.

Plant taxonomy in relation to seed longevity is another interesting field of inquiry. The taxonomic plant families for example, the Cruciferae, the Rosaceae, and the Legummosae commonly are rather uniform in flower structure, type and arrangement of leaves, and other traits.

Are these plant families also uniform in seed biology, or is seed longevity more a matter of ecology ("fit" to the environment) than of taxonomic relationships?

Species of the Leguminosae, considered here in a broad taxonomic sense to include the Mimosaceae, have turned up very frequently in lists of long-lived seeds, and it seems that this family has a marked tendency to longevity of seeds.

Other plant families that apparently have greater than average proportions of species with exceptionally long-lived seeds include the palms, cannas, waterlilies (lotuses), spurges, soapberries, buckthorns, mallows, and morning-glories. Within a single species, varieties and strains with somewhat differing genetic constitution may vary in germination and longevity.

IN CONCLUSION, let us say that we have a great many records some amazing records of long-lived seeds. Up to the present time, however, most data on seed longevity, particularly of wild plants, have been considered piecemeal, species by species.

Much more research and analysis must be done before broad generalizations can accurately relate seed longevity to an integrated consideration of seed and plant biochemistry, physiology, morphology, taxonomy, and ecology.

CLARENCE R. QUICK is a forest pathologist in the Pacific Southwest Forest Experiment Station of the U.S. Forest Service, Berkeley, Calif. Previously he was a forest ecologist in the former Bureau of Entomology and Plant Quarantine.