For these reasons, an accurate and complete formational record that can be consulted or studied at any time is almost indispensable to class study in the development and structure of formations. =244. Details of collecting.= Formational collections, unlike the ordinary sets of exsiccati, can not be made upon the first visit to a region, or by a single journey through it. The determination of formation limits, and of developmental stages, of aspects, layers, abundance, etc., must necessarily precede, a work which alone takes several years. Moreover, collecting itself requires more than one year in a region containing numerous formations. This is exemplified by the _Herbaria Formationum Coloradensium_.[23] The preliminary study for this was made from 1896–1899, the collecting was done chiefly in 1900 and 1901, while additional numbers were added in 1902–3. For the purposes of the formation herbarium, specimens should be collected and pressed in such fashion as to show all the ecological features possible. Plants must be collected both in flower and in fruit, with the underground parts as perfect as may be. Seedlings and rosettes should be included whenever present. In pressing, one or two leaves should be arranged with the lower side uppermost to admit of the ready comparison of both surfaces. Opened flowers are valuable for flower biology, while seeds and fruits are desirable for showing migration contrivances. The ferns, mosses, and lichens of the formation should be fully represented, together with the more important fungi and algae. The number of photographs taken for each herbarium should be limited only by considerations of time and expense. The ideal series consists of a general view of each formation, showing its physiographic setting, nearer views of each of its aspects, detail views of its consocies, societies, and layers, and flower portraits of all the constituent species. Such a series can only be obtained by residence through a long term of years, and in most cases general and aspect views, with portraits of the facies and a few of the striking principal species, must suffice. Quadrat and transect charts, together with formational maps, are extremely desirable, and, indeed, all but indispensable.