Biological Basis for the Breeding of Better Food Yams L. M. Degras Station d'Amelioration des Plantes Institute National de la Recherche Agronomique (INRA) 97179 Petit-Bourg, Guadeloupe The food yams can now be considered to be no more diffi- cult to improve than many other plants. Their system of building up and accumulating useful reserves, their system of ecological adaptation as well as their system of genetic varia- tions offers a sufficient range of diversity to be within the reach of selection pressures. Their sexual behaviour is progress- ively reset among the important species where it was deficient. Evolutive evidence and the sapogenine species experience give wide prospects for attempts of intraspecific introgression. Last but not least, through in vitro technology, new horizons are rising for the modern venture of food yam improvement. Efforts to improve edible yams (Dioscorea sp.p.) are not very ancient. This fact, added to the low number and the wide disper- sal of concerned research workers, account for the doubtful feel- ings often expressed against this improvement. However, it ap- pears not more hazardous than, among others, the banana or the sugarcane ones. The weak international interest in yam exchange is a major difference with these crops and, in turn, accounts wide- ly for their relegation aside of agronomic priorities, while biological hindrances are generally put forth. Our thesis is that basic knowledge exists now, which warrants actual achievements to pluridisciplinary programs aiming at yam improvement through breeding. Biological processes in yams which are at the breeder's disposal can be outlined from three levels: systems of reserve biosynthesis and assimilates accumulation, systems of ecological adaptation, and systems of genetic variation. Considering the genus Dioscorea as a whole, the diversity of biological paths brought to our knowledge offers an uppermost but interesting challenge to the breeding prospects. Systems of Biosynthesis and Storage of Assimilates Photosynthetic Apparatus Even over a small country, the photosynthetic apparatus of the yams can establish itself quite at variation among species and cultivars. The cataphylles phase duration, the phyllotaxy and the vine architecture, the lamina individual area, are concerned. Burkhill (1960) mentioned the wide phyllotaxic spectrum of the genus, going from alternate to seven-verticillate leaves. How such situations could be practically significant can be seen through the low sensitivity to no-staking of the more or less deep- ly lamina indented species (D. dumetorum, D. trifida), accord- ing to the deeper penetration of light in their canopy. Highly different efficiencies of transportation and sink setting of assimilates are to be expected from the different weight ratios of aerial part to tuber, this not only among species but also cultivars. However, care must be taken when the plantation time is changed: its shift from March to June has reversed the rank of ratios of two D. alata cultivars (Degras et al., 1977). Characters of the Products Stored in Tubers There is no need to detail this basic point that a hundred Dioscorea species which store starch (85-99% of tuber d.m.) are edible with a useful content of protein and vitamins, even if for a VOL. XX-PROCEEDINGS of the CARIBBEAN FOOD CROPS SOCIETY limited number a detoxification could be required. It is worth mentioning that: 1. The tuber can be eaten crude in some species like D. fan- dra of Malgasy (Madagascar); 2. In the tuber's enzyme activities noticeable variations have- been found between and within species (Diopoh, and Kamenan, 1981, etc.); 3. In starch viscosity a large variation is known (Coursey, 1980) as well as in flavour at cooking (Osinowo, 1977); and 4. Starch digestibility may be as high (D. dumetorum) as the Cassava one (Szylit et al., 1980), while unknown substances seem to carry some digestibility problems in D. alata cultivars (Martin, 1980). Morphology and Physiology of the Tuber Growth Though a fair number of cultivars need a good soil depth for their tuber, it is not uncommon to access easily to the starch sink. Letting aside the aerial production of D. bulbifera (and at a lower level of some D. alata), rather superficial underground produc- tions are known with D. alata (i.e. 'Lupias' or 'Belep' types of New Caledonia), D. cayenensis (cv. 'Krengle' of Ivory Coast), or D. trifida (cv. 'Moengo 5' of Guyana) etc. Quite different balances between the size of individual tubers and their number by plant can be managed through cultivar variation, enforced by plantation date, vegetative seed size, chemical treatment, plantation densities, staking, fertilization and so on. The well known inter-species variation for harvesting time rests upon a large intra-species one. A recent datum: among a popula- tion of about 400 hybrids pertaining to three progenies of D. cayenensis-rotundata obtained from sexual seeds of IITA in Guadeloupe, maturity extended last year from seven to eleven months after planting. The general inverse correlation between duration of vegetative growth and duration of dormancy over the annual cycle, does not impede some independent combination, as have resulted from selection pressure in D. trifida progenies (Arnolin, unpublished data). A differential sensitivity of cultivars of D. alata, under cool (18 C) storage of their tuber, is expressed by the extension of the onset of germination of 79 cvs from seven to thirteen months after harvest. 93