Because of the similarity of hemp hurds to other materials which have been tested by the Office of Paper-Plant Investigations, semicommercial tests were conducted in cooperation with a paper manufacturer without preliminary laboratory tests. Laboratory pulp and paper tests are regarded only as a preliminary to semicommercial tests and therefore are not employed unless the material in question presents new features which should receive investigation before larger sized tests are undertaken. The advantages of cooperative mill tests are many, among which may be mentioned the counsel and advice of the mill management and employees, the services of specialized and skilled labor, facilities for comparing the processes and the results of tests with commercial processes and results, and the use of commercial or semicommercial types and sizes of machinery. Tests conducted in this manner and on this scale are of a different quality than is possible in those conducted in a laboratory, and the results are susceptible of commercial interpretation with a fair degree of reliability. It is found, in general, that the cost of securing such equipment and service for a complete and comprehensive test does not exceed $500, while the installation of an equally satisfactory equipment alone would cost at least $50,000 and in many cases very much more. Tests conducted in this manner constitute a direct demonstration to the manufacturer, and the results obtained are found to carry more weight when presented to other manufacturers for consideration. It is well known that the method of conducting tests necessarily varies with the size of the test. In the matter of yield determination, for example, laboratory tests may be on such a small scale that the weighing and sampling of the resulting cellulose fibers may be conducted by means of chemical laboratory apparatus and analytical balance, while in tests involving a matter of 5 to 10 pounds of material larger and different types of equipment are necessary. When the tests are so increased in size as to employ 300 or 400 pounds, still other types of equipment are necessary for the treatment of the material and for a determination of the yield of fiber. In tests involving tons of material the equipment involves the use of machines. Accuracy in degree of control and in results will vary materially with the size of the test. As the size of the test increases, certain factors will vary in a beneficial manner, while others will vary in a detrimental manner, so it is a question for each investigator to decide, after taking all factors into consideration, as to the size of test which will give the most satisfactory results. In work of this nature it is found, on the whole, that better results are obtained in large tests, although the control of the factors and the determination of the yield of fiber are more difficult than in smaller tests.