If in some part of North America there had been steady accumulation of sedimentary materials under constantly favorable conditions since the beginning of Cambrian time, the result would have been a deposit of sandstones, claystones, and limestones measuring nearly fifty miles from bottom to top. These figures are based on actual production in North America where extensive measurements have been made in many localities. When other parts of the world are as thoroughly investigated and older deposits included in the calculations, the total thickness of such beds will probably be more than one hundred miles. No single pile of rocks offering a complete cross section of the geological record has ever been produced, but portions of the section are exposed to view on all the continents. In order to carry on desirable investigations and make comparisons, it has been necessary to divide this great composite section into small units which may be named in some way and placed definitely with relation to lower and higher, or older and younger, layers. To serve this purpose there has been developed the idea of rock _formations_, and here we have a word which is not defined readily, even for the use of those who are familiar with it. Nevertheless it is used so commonly that some understanding of its meaning becomes desirable. A _formation_ may be regarded as an extensive rock mass, variable, in thickness and other proportions, as well as in composition, but representing a period of time during which there was no great change in the character of plant and animal life, and no serious interruption in the depositing of the rock-making materials. Occasionally the lower and upper limits of a formation are well defined and readily located. Frequently, however, the transition is gradual, one formation merging into another with no apparent mark of separation. In such event the original description serves to establish more or less definitely the boundaries of a formation. Descriptions are published whenever a worker believes he has discovered a significant part of the great section which has not previously been named. The usual practice is to apply a name taken from the locality in which the beds were investigated, and in this manner the names of formations become associated with towns, rivers, counties, mountains, states and other geographical features. The locality which supplies the name is then regarded as the “type locality” for the formation, but wherever these same beds may be traced or otherwise identified the one formation name applies. [Illustration: Dinosaur Tracks An ancient trail in sandstone of the Dakota formation. East slope of the hogback, west of Denver.]