sample areas were established in an undisturbed part of each
 slash pine plot. These small sample plots were located in the
 center of a square formed by four adjacent trees. The surface
 litter was carefully raked away, and a small trickle of water
 from a garden hose was allowed to flow onto the sample area.
 Only one quarter of each 0.184 square meter plot was sampled
 at any one time. A jacknife was used to loosen the soil and to
 collect all pine roots less than 0.5 cm diameter to a depth of
 15 cm. Progress downward was at the rate of a few millimeters
 at a time. This method was very laborious but resulted in a
 minimum loss of fine roots.

 Laboratory
 All tree components were dried to a constant weight of 65 C,
 and the dry weight was recorded. Discs were cut from the top
 of each bolt, redried, and weighed with bark on. The bark was
 then removed, and the discs were redried and reweighed. A
 ratio was used to calculate the weights of bolewood and bole-
 bark for each tree.
 Large roots (more than 0.5 cm in diameter) were separated
 into lateral roots and taproots. Fine roots were separated from
 soil particles by washing and decanting through wire screens
Sand cheese cloth. Corrections in fine root biomass for soil par-
 ticles not removed in washings were made on the basis of ash
 weights.
 Branches, bolewood, and large root components were chipped
 in a hammer mill and subsampled for further treatment. All
 components were ground separately in a Wiley mill to pass a
 40-mesh sieve.
 Subsamples were analyzed for N by the macro-Kjeldahl pro-
 cedure (3, 6). Other subsamples were ashed overnight at 450 C,
 and the ash was dissolved in dilute HCI for analysis. Aliquots
 of these solutions were analyzed for P by the molybdophosphoric
 blue method (6) ; K by a Beckman DU flame spectrophotometer
 (6) ; Mg, Ca, Al, Mn, and Zn by a Perkin-Elmer 303 atomic
 absorption spectrophotometer. Lanthanum oxide was used in
 the Mg and Ca determinations to supress interference (1). .
 Soil total P was determined by heating at 520 C for 4 hours
 and then digesting with 12 N HC1 on hot plate, according to
 procedure of Murphy and Riley (7).
 Standard statistical methods have been used to analyze the
 field and laboratory results (2, 12).