GEOGRAPHY OF CENTRAL FLORIDA 193

stituent appear somewhat contradictory, for some of the samples in the last table show more of it than one could reasonably expect. Taking everything into consideration, however, it is safe to say that central Florida soils generally contain less potash than those of northern Florida or any equal area a few hundred miles farther inland. The reason for this is probably two-fold: first, the remote.ness of this area from igneous rocks which are the main original source of potash, and second, the leaching effect of the copious late summer rains. But this lack is partly compensated by the temperature, for the plant food in any soil is liberated more rapidly in a warm climate than in a cold one.
 The scrub and dune, soils are low in potash, as in almost everything else.
 The lire., like the potash, is as a rule most abundant in the richest soils, and vice versa, but there are some important exceptions. For example, the beach and dune sands are well supplied with calcium carbonate in the form of shell fragments, but are practically worthless for agricultural purposes, perhaps chiefly on account of the scarcity of very fine particles for plant roots to draw nourishment from (no mechanical analyses seem to be available, Unfortunately), or of soil animals and bacteria. And the vegetation of the calcareous hammocks near Ocala (sample T), with over 7% of lime in the soil, does not seem as luxuriant as that near McIntosh (sample U), where there is less than 2%. However; probably the latter figure is more than sufficient, and any excess over that therefore superfluous. The least lime is in the St. John's River prairie (Y), which seems rather strange, for ditches near where the sample was taken show shell marl within two or three feet of the surface.
 Although lime (or more strictly speaking any calcium compound) is not an important plant food, it is thought to improve the condition of the soil in various ways, and as it dissolves readily it liberates less soluble plant foods that may be combined or mixed with it.
 The phosphorus is almost incredibly high in samples R, T, U and X, soils with more than half of one per cent of P20 being very exceptional.* The high percentage in X, which was taken

 *See Hilgard, Tenth Census 5:78; 1884; Soils 355. 1906.