OTHER THINGS being equal, a large sample is more representative of a seed lot than a small sample. A sample must be large enough to permit a test and retest.

The following amounts usually are sufficient: Two ounces of the fine grasses, whiteclover, and all seeds of similar size; 5 ounces of red clover, lespedeza, bromegrass, flax, and seeds of similar size; 1 pound of sudangrass, sorghums, proso, and seeds of similar Size; 2.5 pounds of cereal, vetch, and the larger seeds; 800 seeds of vegetable seeds, which require a germination test only.

Additional seeds must be available if varietal trials are to be conducted in the field or greenhouse.

If these sampling procedures are followed, the bulk sample submitted to the laboratory will be larger than the amount needed for testing. The method used to obtain the working sample from the bulk sample submitted will determine how well the test results describe the seed under consideration.

The International Seed Testing Association studied several methods of subdividing samples and found the use of a good mechanical divider is better than the halving method.

In the halving method, the sample is placed on a clean piece of paper or oilcloth, mixed with a spatula and rolled on the paper or cloth, and marked into quarters with a spatula. Opposite quarters are discarded. This process is followed until the final sample is about the size required for analysis. This method theoretically should yield representative subdivisions, but in practice it does not. It is not recommended for field or laboratory use.

Mechanical dividers produce repeated working samples that are about as uniform as one can expect from true random samples.

Several kinds of mechanical dividers are available. The Boerner type is used widely. It is available for small and large seeds. Any riffle-type mechanical divider is superior to hand methods for reducing the bulk sample to the working sample.

Some State seed laws require that the inspector divide the sample as soon as it is drawn and leave a portion with the retailer. This provision is satisfactory if a good mechanical divider is used to divide the sample, but the shortcomings of hand division make the use of this method in the field or in the laboratory questionable.

The warehouse is a poor place to subdivide a sample. Warehouses are apt to be cold or hot or drafty not good places in which to subdivide seed samples. It is much better, when the rules allow it, to take the entire sample to the laboratory where mechanical dividers are available.

THE PERSON who does the sampling needs to be trained in the methods of obtaining a true sample. He should have proper equipment.

In drawing a composite sample to determine the quality of a lot, the labeling on each bag sampled must be examined to be sure that all bags belong to the same lot. Each trierful should be examined to detect noticeable variation before combining. If variations from bag to bag are noted, individual bag samples should be drawn to determine the extent of the lack of uniformity in the lot. When the sample is drawn, the sampler must identify the containers into which it is placed before he draws another sample. He must be careful that the sample is not contaminated with dirt or other material.

The identity of each sample must be maintained throughout the sampling and distribution process. Most official samplers put an inspection slip inside the container with the seed and also identify it by writing a description on the outside of the package. Records accompanying the sample must be complete so there will be no difficulty in associating the final report with the "definite quantity of seed" that has been sampled.

Some farmers prefer to sell seed at wholesale and let the seedsmen worry about testing and labeling it. Others prefer to process and market the seed they produce.

Because most State laws require farmers to label seed offered for retail sale, farmers often are required to draw samples of seed lots and submit them to commercial or State laboratories for testing. Other farmers get their homegrown seed tested to determine its planting value.

Farmers usually do not have triers and mechanical dividers and other equipment. They can draw a fairly representative sample, however, by taking handfuls of seed from different places in different bags in the lot, or in different parts of the bulk lot, mixing this seed together, and taking handfuls of the composite. This sample will not be so accurate as one drawn with an approved trier by a professional inspector, but it will be acceptable.

Farmers may fail to submit satisfactory samples because they sample only one bag or too few bags or too few places in each bag; their samples may be too small for a complete test; samples of bins of seed often are taken from the place most easily reached, instead of from different places and depths in the bin.

Farmers often do not identify samples properly as to lot number or other designation when they send them to the laboratory.

So the sample is all-important. It should be drawn by a trained person using good equipment and procedures. It should be subsequently subdivided with the best mechanical divider available and with utmost care. Commercial samplers and official inspectors should always be warned that the sampler of seed has someone's reputation and well-being in his two hands.

A. S. CARTER, Director of Seed Control and State Chemist Services, Biochemistry Department, Purdue University, is concerned with the administration of the Indiana seed, feed, and fertilizer laws. He has conducted research in seed technology, including the sampling of seeds.

A seed mixer and divider, which is used to reduce a laboratory bulk sample to a working-sample size.