=43. Taking samples of soil.= In obtaining soil samples, the usual practice is to remove the air-dried surface, noting its depth, and to sink the geotome with a slow, gentle, boring movement, in order to avoid packing the soil. This difficulty is further obviated by deep notches with sharp, beveled edges which are cut at the lower end. In obtaining a fifteen-inch core, there is also less compression if it be cut five inches at a time. Repeated tests have shown, however, that the single compressed sample is practically as trustworthy as the one made in sections. The water-content of the former constantly fell within .5 per cent of that of the latter, and both varied less than 1 per cent from the dug sample used as a check. As soon as dug, the core is pressed out of the geotome by the plunger directly into an air-tight soil can. Bottles may be used as containers, but tin cans are lighter and more durable. Aluminum cans have been devised for this purpose, but on account of the expense, “Antikamnia” cans have been used instead. These are tested, and those that are not water-tight are rejected, although it has been found that, even in these, ordinary soils do not lose an appreciable amount of water in twenty-four hours. The lid should be screwed on as quickly as possible, and, as an added precaution, the cans are kept in a close case until they have been weighed. The cans are numbered consecutively on both lid and side in such a way that the number may be read at a glance. The numbers are painted, as a label wears off too rapidly, and scratched numbers are not quickly discerned. [Illustration: Fig. 4. Field balance.]