soils or in ecological conditions that are not today considered
favorable for sweet potato. Where feasible, we also wish to im-
prove the production systems, whether in mono- or multiple-
cropping, in order to reduce risk to farmers. Finally, the objective
is to broaden the use of sweet potatoes so that the impact of the
crop is realized.
 In order to accomplish these goals we are breeding sweet potato
by mass selection. Our mass selection program requires a full year
for each cycle. We plant seeds in January to April, transplant
them to the field, screen for characteristics desired either in field
or greenhouse, and then select plants to be used, together with
the best of previous selections, in poly-crosses that will flower and
bear seed during the dry season.
 Getting a good sample of germplasm was our problem at the
beginning. Now, having introduced about 40,000 seeds from the
USA, West Africa, the French West Indies, Central America, and
the South Pacific, we have all of the germplasm that we need.
Please note that the clonal varieties of sweet potato cannot be
legally imported into the USA or Puerto Rico, unless they pass
through rigid quarantine procedures at Glenn Dale, Maryland.
Within the germplasm available there is a broad scope for selec-
tion, and new variation is readily produced following hybridiza-
tion. Germplasm improvement is accomplished through up to 15
poly-crosses each year. We now have seeds to spare for other in-
vestigators.
 Closely related to the question of introduction of germplasm,
is that of the evaluation of germplasm already here in Puerto
Rico. In fact, this emphasis cannot be separated from the testing
and evaluation of new varieties of our own program In order to
evaluate varieties we had to develop standards. We did this for
what we consider four types (Table 4). The dessert type is that
most appreciated in the southern USA, orange, moist and very
sweet. The tropical type, ranging from white to yellow, from low
to high in sweetness, and from intermediate to dry in the mouth,
is already found throughout the Caribbean. The staple type is
what we are now beginning to produce, white, without
sweetness, hopefully neither too dry nor too moist, and with a
bland flavor suitable for everyday use. The high nutrition type,
which could contain large amounts of vitamin A in a non-sweet
type, is still only a dream, although we have seen prototypes.
 We have finished a series of tests for our first ten varietal selec-
tions of the tropical and dessert types and have seen that some are
well adapted, even to heavy, poorly drained soils and acidic con-
ditions. At this early stage of our breeding program, I already
believe we have significant new varieties. However, we have just
begun to test a second set of varieties. These are our staple and
substaple varieties, and I am confident that they are going to be
better than the first set, at least in table quality.
 By developing our sweet potatoes in heavy, acidic, poorly
drained, infertile soils, we are subjecting them to stresses of
several kinds. The Mayaguez soils (Catalina clay, Cialitos clay) are
not very well suited to sweet potatoes, but they are representative
of the tropics. We have seen that sweet potatoes that produce
well in these soils, also tend to produce well in other soils. On the
other hand, varieties that are well adapted to, and productive in
other soils are not necessarily adapted to Mayaguez soil. Thus,
our aim to breed stress-resistant sweet potatoes is met in part by
breeding in the Mayaguez soil where the most important stresses
seem to be high acidity, root flooding, and low fertility.
 However, breeding for stress resistance also involves the use of
greenhouse tests where large numbers of plants can be screened
in a small space. Since sweet potatoes will produce a small storage
root even when cuttings are planted in 3-inch pots, stress factors
impeding production can be studied in the greenhouse. The
stress tests that we have developed and standardized include root
flooding, shading, soil acidity, poor fertility, drought, and com-
petition. With each of these tests, we are finding that seedlings


vary in their reaction to stress. We test single cuttings of seedlings
and if stress tolerances are found, we test again on a triplicate
basis. Stress resistant seedlings then are poly-crossed to produce
the next generation.
 What we do not yet know is whether the stresses imposed in an
artificial situation, the greenhouse, bear any relationship to
stresses in the field; yet, we suspect that they do. In our first selec-
tion for shade resistance, all 38 selections out yielded all 18 stan-
dard varieties in a field plot beneath trees. Furthermore, in a test
of 100 seedlings selected at random, subjected to greenhouse
tests but also grown in three situations in the field, stress test
results, especially for tolerance of acidity, were partially correlated
with yield performance in the field. In addition, using only stress
test results and selecting the top 15 % of the seedlings, we would
have been able to select nine out of the top ten yielders in
Mayaguez fields.
 With this kind of research, we are trying not only to develop
new sweet potatoes, but also to determine something about their
ability to adapt to other environments different from those of
Mayaguez.
 As a prominent factor influencing yields, we have identified
the stem cutting and its treatment before and after planting as
being of crucial importance in yield. For example, healthy-
appearing cuttings from old plantings do not produce as well as
similar cuttings from younger and more vigorous plantings. The
nutritional status of the cutting can range from dismal to promis-
ing. Furthermore, the cutting may suffer from distinct levels of
virus charge. More exact information on the influence of pre- and
post-planting procedures will permit the development of better
sweet potato production systems.
 In the tropics the sweet potato is frequently produced in multi-
ple cropping systems. Our answer to the special needs these
systems require is the development of climbing sweet potatoes to
be combined with corn, and development of shade tolerant sweet
potatoes to grow under trees such as coconuts. We have rapidly
produced climbing varieties using selection and poly-crossing,
but these sweet potatoes are different and require much study
before they can be released. For example, they have small stems
and are difficult to root. They may be more sensitive to
photoperiod, and may accumulate virus charge more readily.
Thus, breeding has been very successful in producing a new kind
of plant, but one which we do not know yet how to handle.
 The search for shade tolerance is not difficult, but results so far
suggest that the level of tolerance is low. It may take many years
of mass selection to develop a truly useful level of shade
tolerance.
 I would like to be able to report progress with respect to the
sweet potato weevil, but like all farmers who know this pest, I
have no sweet words. We think we have a good methodology for
searching for weevil resistance so that it can be strengthened by
mass selection. Now, with almost four years of field experience
behind us, we feel that weevil resistance is uncommon and weak,
thus hard to find, and harder yet to strengthen. Yet we feel that
the search for weevil resistance is very necessary, and that the
germplasm under development may as yet be the most important
contribution we can make. We urge farmers to use those few con-
trols that can permit reasonable production in spite of weevils.
 I have not touched on the subject of processing, but wish to
mention two contributions we have made. One is the develop-
ment of diffusion-processed sweet potato pulp through a simple
process that makes any sweet potato taste better. With this pro-
cess sweet potato could be used as a staple food in any household.
The second contribution is the development of a precooked, non-
sweet flour from one variety of sweet potato that would permit

excess from the harvest to be stored for future use. In order to
develop a process suitable for small scale use, we have had to
develop technical solutions suitable for the average household. It


PROCEEDINGS of the CARIBBEAN FOOD CROPS SOCIETY-VOL. XX


204