TABLE 1. Selected characteristics of Virgin Islands farming: 1960 to 1982.


 Farm Units
 Produc- Z Farms
 Growth tive Producing
 Year No. Z Change Factor Acreage for Sale

 1960 501 44062 60.3
 1964 466 -7.0 0.93 39539 49.1
 1970 212 -54.5 0.45 20470 48.6
 1975 327 54.2 1.54 24703 54.1
 1978 378 15.6 1.16 24397 46.6
 1982 303 -19.8 0.80 20824 71.6

 Average
 rate of
 change -10.6% -9.6% -16,2%

 Note. The data in column 2 are respectively from the U.S.
 Census of Agriculture, Virgin Islands of the United States,
 U.S. Bureau of the Census.


which there was a combined total of ten or more fruit or nut trees
or plants, any livestock, or ten or more poultry at the time of
enumeration. Places of less than three acres were counted as farms
if their sales of agricultural products amounted to at least
$100, or if they could normally be expected to produce
agricultural products in sufficient quantity to provide sales
amounting to at least $100 ... The definition of a farm for 1982
was the same as in all previous census, except the one for 1950"
(U.S. Bureau of the Census, 1983, p. iv). In fact, Table 1 shows
that the total number of farms in 1964, 466, was reduced to less
than half in 1970 when only 212 remained. And concomitantly,
the total productive acreage recorded the largest decline in the
last 25 years, a change of 52.7% or a growth factor of 0.473.
 The pattern of change that is reflected by the data of Table 1 is
one of decline, but this has not been consistent from one quin-
quennium to the next. First, it is noted that overall there is a
10.6% rate of decrease per period between 1960 and 1982, in the
number of farm units in the territory. The table also shows that
the decline has not been constant: the third and fourth columns
of percent changes and growth factors respectively express uneven
rates of decrease and increase. The average period-to-period per-
cent change is therefore 9.6%. The acreage devoted to farming
is related to the number of units under production, and the data
thus reflect the same type of increase and decrease. Overall, the
average rate of decrease between periods is 16.2%.
 The most recent census records the largest percentage since
1960 of the total number of farms that produce for sale, versus
those that produce for home consumption only. Whereas about
six out of ten farms were producing for sale in 1960, the rate in
1982 was seven out of ten.

 Census documents group most of the farm data by size
categories. It is therefore useful to examine, by comparison, the
frequency distributions of the last five quinquennial censuses to
determine if the number of farm units in each of the categories
remains basically the same from one census year to the next, or if
they increase or decrease. The last decennial census of agriculture
in the Virgin Islands was taken in 1960. The first quinquennial
census of agriculture was taken in 1964. For the 1969 quinquen-
nial census, the data were collected in 1970 to coincide with the
decennial Census of Population and Housing. In 1976 Congress
authorized censuses for 1978 and 1982 "to adjust the data
reference year to coincide with the 1982 Economic Censuses .
after 1982, the agriculture census will revert to a 5-year cycle"
(U.S. Bureau of the Census, 1983, p. iv.). The data in Table 2 il-
lustrate that, whereas the percentages of the number of farms


 TABLE 2. Percentage of Virgin Islands farms by size: 1964 to 1982.


 Size of farms
 (in acres) 1964 1970 1975 1978 1982

 Less than 3 10.3 31.1 33.3 24.1 24.8
 3 to 9 42.0 25.9 28.8 38.9 37.6
 10 to 19 16.1 11.3 10.7 14.3 10.6
 20 to 49 10.5 11.8 10.1 9.0 11.2
 50 to 99 6.8 7.6 5.8 5.0 5.9
 100 to 174 5.2 3.3 4.0 2.1 4.0
 175 to 259 2.2 1.4 2.1 2.1 1.6
 260 to 499 3.0 3.3 2.5 2.4 2.0
 500 to 999 2.4 1.9 0.9 0.8 1.3
 1000 or over 1.5 2.4 1.8 1.3 1.0

 100.0% 100.0% 100.0% 100.0% 100.0%
 (466) (212) (327) (378) (303)


over 100 acres in size have remained rather constant, the numbers
in the smallest category (of less than three acres) have varied
remarkably. Among the number of farms over 1,000 acres, there
were 1.5% in 1964, and 1.0% in 1982. In 1964, 10.3% of the
farms were less than three acres in size, and this number tripled
to 31.1% by 1970. The largest percent of 33.3 was recorded in
1975, and the most recent count listed 24.8%. No other category
registered this degree of variation in the proportion of farms in
intercensal periods. The large increases between 1964 and 1975
may probably be explained by the dismemberment of some of
the sugar plantations that ceased sugar cane production in 1966.
 It was of further interest to examine the five distributions to
determine statistically if it can reasonably be said that, basically,
the same distribution of the percentage of farms was maintained
in each of the census years from 1964-1982. The chi-square test of
homogeneity is utilized to establish if the differences between the
distributions can be ascribed to a chance process. The validity of a
statistical test of significance on population data may have crossed
the reader's mind. However, the attempt is not to generalize here
to a larger population; it is merely to rule out the 'chance process'
alternative as an explanation for the observed differences. See
Blalock (1979, p. 242). The computed chi-square and associated
probability of p = .00011 suggest that the populations are
neither identical nor homogeneous, and that chance is not
responsible for the observed differences. In fact, the test confirms
that the largest differences are between 1964 and 1970, and these
may be attributed to the agrarian change in sugar production that
characterized the period.
 Commensurate with the decline in farm unit numbers was a
decrease in the number of acres devoted to productive
agricultural practices. Table 1 recorded a quinquennial decrease
of 16.2% over the period, and Table 3 illustrates the percentage
of agrarian land that was distributed among the various size
categories. Typical of most farming systems throughout the
Caribbean, less than 1% of the total farm acreage in 1982 was
found on 25 % of the farms under three acres, and for most years,
more than half of the total agrarian land was confined to farms
over 1,000 acres in size. While the percentage of the acreage on
farms of less than three acres tripled from 1964 to 1970, i.e., from
0.2% to 0.6%, the actual amounts were 69 (1964) and 117 acres
(1982). However, the percentage in almost all categories from 3
to 999 acres decreased between 1964 and 1970. The category that
registered an increase was farms over 1000 acres: from 39.2% in
1964 to 54.7% in 1970. The last census marks the first time in
about 15 years that the percentage of farms in the largest category
is less than 50%, 45.7% in fact.


PROCEEDINGS of the CARIBBEAN FOOD CROPS SOCIETY-VOL. XX


24