Ulotli on tho root of a maple seedling which penetrated a short distance into solid ice. Wheat has been known to germinate at the freezing point.

The following table, showing the effects of given temperatures upon the germination of seeds, is taken from Nobbe's Handbuch der Samenkunde. The column under a indicates the number of seeds germinated; that under b shows tho number of hours required to germinate that number under the fixed temperature.

PROPER CONDITIONS FOR TESTING SEEDS.

The best temperature for the germination of most seeds is shown to be 25° C. (77° F.), while for a few this optimum is 31° C. (88° F.) and 37.5° C. (100° F.). But seeds germinating under natural conditions seldom have the advantage of this optimum temperature.

In testing seeds, therefore, since it is necessary to get as near the natural conditions as possible, the temperature should be kept at between 18° and 20° C. (64° and 68° F.). This has been found to be the normal temperature for germination. Usually tho heat of an ordinary living room will be sufficient for home testing, but if the temx)erature is likely to fall very low during the night it is better to provide a little heat during that time. More harm will result from a considerable decrease of temperature than from a slight increase. In the European seed-control stations seeds are tested at a constant temperature of 18° to 20° C. (64° to G8° F.). For grass seeds the temperature is forced up to 30° C. (80° F.) during six hours of the twenty-four, this variation in the boat being found advantageous.

Moisture is as important as temperature. Before a seed can sprout it must absorb water and swell. Though the swelling of a seed is a necessary preliminary, it is not always followed by germination, for the absorption of water is a purely mechanical process and does not imply vitality in the seed. The entrance of water into the seed is dependent upon the structure of the seed coats. When these are hard and impervious, as is often the case in leguminous seeds and in nuts, water gains admission slowly and germination is retarded. In cereals and in most garden seeds the seed coats are easily penetrated by water, the seeds swell rapidly, and germination is prompt. Experiments have proved that seeds will absorb moisture and swell in a damp atmosphere, but that for germination, contact with water is necessary. An atmosphere saturated with water vapor is not sufficient to induce germination. Flaxseed kept in a saturated atmosphere for nine days, and seed of kohl-rabi kept under the same conditions for twenty-two days, did not germinate (Nobbe, Handbuch der Sameukunde). Too much water is equally injurious. As a general rule, seeds will not germinate well when immersed in water. It is necessary to have the seeds in contact with some medium from which they can obtain an abundant supply without allowing water to stand around them.

Light exerts a harmful influence upon germination. Experiments have shown that seeds placed under colored glass did not germinate as rapidly as those which were in complete darkness. Even more important than the exclusion of light is the free access of air and the escape of the noxious gases generated by germinating seeds. When germination has commenced, carbonic acid gas is given off, which must bo allowed to escape, or growth will be checked.

SELECTING SAMPLES.

Selecting the sample to be tested is a matter of great importance. It must be a fair sample, including both good and bad seeds. If the quantity to be tested is considerable, small amounts should be taken from different parts of the mass. These small samples, thoroughly mixed, form the larger sample out of which the proper number of seeds is to be counted. In case the quantity of seed is small, say one-half pound of clover seed, pour the seed from the package into a pan, taking a small spoonful occasionally from the stream. From the quantity thus secured a sample for testing is taken. The number of seeds used in testing depends upon the size of the seed and upon the quantity at disposal.

If the sample is largo enough, 100 seeds of tho larger kinds and 200 to 400 of the smaller seeds are taken. The increased number is a check upon error in counting small seeds. In counting out the seeds a fair number of small and immature ones should be selected as well as the large and plump ones. There is reason to suspect that in some tests only fine-looking seeds arc used. These would, of course, give a higher percentage of germination than could be sustained by the entire sample. In selecting grass seeds for testing, care must be taken to use only such as contain a grain. In some kinds of grass seeds there are many empty glumes which it is difficult to distinguish from those containing a grain. A simple way to separate them is to wet the seed, spread it out on a plate of glass, and hold the plate up to the light. The empty chaff will appear translucent, while the good seed will be opaque.

KEEPING A RECORD.

Although for the results usually desired in home seed testing it is not absolutely necessary to keep a record, yet such a record, if well made, will be found to contain much valuable information. A few items will always need to be recorded, in any event, such as the date of beginning the test, the name of the variety, the number of seeds, and the number of germinated seeds removed from day to day. It is dangerous to trust anything to memory. Mistakes are sure to occur, ami the test will then be useless.

LENGTH OP TIME REQUIRED.

The length of time a test should continue depends upon the seed. In the seed-control stations ten days has been accepted as the proper time for most seeds, but a few require a longer period, namely:

Days.

Esparsette, serradella. 1 eet-seed balls, rye grasses, timothy, carrots 14

Grasses, except meadow and rye grasses, and timothy. _ 21

Meadow grasses (Poa), conifene (except white pine), birches, alders, acorns,

be- ches, and hornbeams •_ 28

White pine and stone fruits 42

The seeds should be examined each day, and those that have germinated should be removed and the number recoi-ded. A seed is considered as germinated as soon as the root breaks through the seed coats.

Under favorable conditions more than one-half of the seeds in a good sample will germinate in a much shorter time than that given above. The rapidity with which the seeds germinate is some indication of the vigor of the embryo, and determines the germinative energy.

The number of days in which more than one-half of the seeds in a good sample should germinate has been fixed as follows:

Days.

Cereals, clovers, peas, vetches, flat peas, flax, dodder, poppy, cabbage, radish, spnrry, chicory 8

Squashes and pumpkins, encumbers, beans, spinach, lupine, buckwheat, burnet 4

Beet, timothy, serradella, bird's-foot clover, rye grasses, meadow foxtail, reed grass 5 Days.

Redtop, hair grass, chervil, carrots, fennel, esparsette, sorghum 6

Spruce, fox-tail grass, sweet vernal grass, canary grass, Deschampsia, Trise

tum, Poa, crested dog's tail, velvet grass, red and sheep's fescue 7

Fir, pines (except white pine), maple 10

White pine' - - 14

In nearly every test, especially of leguminous seeds, there will be some that remain hard. These can not be regarded as dead seeds, because their condition is due to the hardness of the seed coats. The number of such seeds should be recorded.

SPECIAL CARE NEEDED IN TESTING BEET-SEED BALLS.

In testing beet-seed balls special care is necessary in recording the number of germinated seeds. The balls must be left in the test

Fig. 23.—Simple germinating apparatus. A, closed; B, open.

during the entire period of fourteen days, but whenever a seed has sprouted it must be cut out with a sharp knife; or the root may be allowed to grow two or three days and then broken off and counted. The roots will either not grow out again, or, if they do, can not be mistaken for fresh ones. Either operation is very simple, and can be done by any one without the least trouble. The removal of the germinated seed or of the young roots is the only sure way of making an accurate test of the germination of beet-seed balls. One hundred seed balls should produce at least 150 seedlings.

1 Yearbook, U. S. Department of Agriculture, 1804, p. 399.

APPARATUS.

The apparatus used for home seed testing should be as simple as is consistent with a reasonable degree of accuracy. Any method that complies with the conditions given above—a proper amount of heat, moisture, air, and the exclusion of light—will give good results. Fortunately, these conditions are so easily fulfilled that the most inexpensive apparatus will answer. Perhaps the simplest and at the same timo the most satisfactory is the following:

Take two plates and place in one of them a folded cloth; wool or flannel is preferable, since it remains moist for a long time, but any cloth will do. The cloth should be free from dyes that will come out in water, since they may contain chemicals that would bo injurious

Fio. 21.—Homemade germinating apparatus. A, complete; B, section.

to the seed. Wet the cloth, pressing out the surplus water, leaving it very damp, but not soaked. Place the seeds between the folds of cloth, put in the number of the record, marked in pencil on a piece of paper, with date and number of seeds, and cover with the second plate, inverted. Plenty of air will get in between the plates, and the upper one will prevent too rapid evaporation of moisture. If the tests are to be made during the winter, keep the apparatus in the living room, as the heat of such a room will be sufficient for most seeds. During the night the seeds should be put in a warm place. Instead of the cloth, old newspapers, well soaked, can be used. These need to be moistened more frequently, however. (See fig. 23.)

Another apparatus that will give good results, especially for seeds not larger than wheat, is the one shown in fig. 24. Here the seeds are placed free on the bottom of a porous saucer and the latter put inside of a tin basin. The basin should have at least two coats of