The yields of traditional systems, which are significant and
relatively regular, rest therefore on a number of mechanisms,
some of them playing on opposite directions (like "maximal time
occupation" and "late cultivars"), perhaps with a better
homeostatic result. All these mechanisms, and their relative im-
portance, are not well known today, for example, nitrogen fixa-
tion by Azospirillum on cereal roots, or endomycorhiza on cassava
or legumes. Most of them probably do not work with maximal ef-
ficiency:
 1. For nutrients, especially on ferrallitic soils, recycling of
 vegetable by-products, even from deep-rooted plants, will
 not be enough to correct major or minor element deficien-
 cies (phosphorus, molybdenum).
 2. For cultivars, inter- and intraspecific heterogeneity allow
 individual plants, either resistant but low yielders, or
 higher yielders but susceptible, or even without any quali-
 ty, to survive in the mixture.
 There would be, therefore, good reason to elaborate
ameliorated systems, taking advantage of what can be learned
from scientific studies on the biological mechanisms operating in
the traditional systems, without declining either a reasonable
amount of chemical aid, or adapted plant breeding.

 For this objective research we need to follow a number of
paths.
 1. Better knowledge of biological mechanisms of soil fertility
 maintenance; evaluation of restitution by stems and leaves
 of annual plants or dead leaves of trees and shrubs; better
 knowledge of the systems of nitrogen fixation and soil and
 subsoil mineral exploitation other than the classical
 Rhizobium one (bacteria associated with cereal and grass
 roots, endomycorhizae);
 3. A general policy of restriction of system losses: better use of
 vegetable and animal wastes, and ability to stimulate the
 system efficiency with minimal inputs of phosphoric acid,
 calcium, magnesium, potassium, etc.
 4. Definition of a number of crop associations or successions
 with proved reciprocal benefits, the partners of which
 could be introduced in plant breeding programs, either for


 their food productivity, or for their soil fertility restoration
 performances, or for both. Plant breeding is indeed a very
 efficient system, which gives solutions it was asked for, for
 example, "varieties tolerating and utilizing large amounts
 of chemical fertilizers." Other solutions can perhaps be ob-
 tained: varieites with wide adaptation, which would be the
 best both in medium-fertile and poor soils, varieties
 ameliorated both for yield and fertility restoration
 (perhaps for pigeon peas), or even plants bred only for soil
 fertility restoration.

Results in Coincidence
with the Orientations Defined Above
 The first result was obtained by the French cooperators in
Haiti: it concerns a very simple traditional system, growing
Phaseolus vulgaris beans in pure crop, but with varietal mixture,
alternating with 9 or 21 months fallow. The use of a bean line,
tolerant to powdery mildew and rust, and of a minimal 0-30-40
fertilization, obtained with wood ashes and superphosphate,
allows the harvest of 1.5 t dry beans/ha in the place of 0.5.
 For the French West Indies, we begin to have in hand the
varietal components of ameliorated creole gardens: for example,
L. Degras' collection of alata yam clones tolerant to anthracnose;
lima bean lines which, although non-photoperiodic, are able to
survive more than one year without pesticides or nitrogen; the
best cucumber lines of American catalogs, resistant to downy and
powdery mildews; and, before long, the best sweet potato clones
bred by Dr. F.W. Martin.
 The best associations of such cultivars are still to be chosen, and
a minimal fertilization scheme to be found. The study of the
thermophilic microflora of composts in tropical conditions was
begun this year by S. Peltier. It is somewhat similar to the flora
described in temperate countries, with Actinomycetes, Bacillus
spp. and fungi. Some of these microorganisms seem to be an-
tagonistic towards Sclerotium rolfsii and the tannia decline
Pythium. There is also a possibility of colonization, at lower
temperatures, by Trchoderma spp. on organic substrates in order
to control Sclerotium rolfsii. The work has begun; general syn-
thesis still remains to be made.


VOL. XX-PROCEEDINGS of the CARIBBEAN FOOD CROPS SOCIETY


221