Florida Agricultural Experiment Stations


composition of leaf and fruit analysis indicated that smaller vari-
ations were found in the potassium and magnesium, but some-
what larger variations in the nitrogen, phosphorus and calcium
contents of fruit samples.
 Mineral Composition of Soil.-Soil samples from the 0-6 inch
depth were analyzed for available phosphorus, extractable cal-
cium and magnesium, copper and soil reaction (pH). In gen-
eral, the variations in phosphorus, calcium and magnesium con-
tents covered a much wider range in soil than in either leaf or
fruit. It should be pointed out that the data reported represent
only the quantity of these elements extracted by certain acid
solutions. It is not the total quantity nor does it tell what frac-
tion was extracted.
 Copper content in soil is expressed as pounds per acre, with
50 pounds considered as low, 100 pounds as medium and 200
pounds as high or excessive. The average and standard devia-
tion for copper were not calculated in Table 12. Because of
the manner of expressing the data, such calculations will serve
no useful purpose. The soil reaction of the groves under study
covered a range from pH 4.8 to 6.8. Approximately 85 percent
of the groves were in a pH range of 5.4 to 6.2, which is recom-
mended for growing citrus on acid sandy soil (12).

 DISCUSSION
 The following discussion represents the authors' interpreta-
tions of some of the trends observed. It should be pointed out
that in any survey type of study, where conditions are not con-
trolled, wide variations can be expected among the individual
cases, related not only to the variables studied but also to other
variables for which no measure may be available or which are
unrecognized. The lack of control of the variables and the non-
uniformity of unmeasured factors may obscure the true rela-
tions, so that much caution is necessary in interpreting the re-
sults. Conclusions drawn from the present study are subject to
these reservations, and should not be mistaken for the more pre-
cise results of controlled experiments. It is felt that enough
groves were included in the present survey to be representative
of grove conditions in the State, and to allow the detection of
obviously significant trends. Some of the variables, such as
rootstock and variety and to a lesser extent soil type and tree
age, were elimniated during the selection of groves.