Introduction:

The chicozapote is an increasingly important fruit-tree crop in India, Mexico, and other tropical regions of the world. Over 23,000 ha of land in India are devoted to production of this delicious fruit, a tenfold increase from just twenty years ago (Chadha 1992). It thrives on a wide variety of soils and water regimes, and is resistant to salinity, drought, insects and pathogens. In the Yucatan peninsula of Mexico, the tree often occurs in morphologically distinct populations in forests and periodically inundated swamps. The populations differ greatly in plant height, architecture, fruit size and seed size (Heaton, unpubl. data). Swamp populations are much smaller in size than their forest counterparts and have smaller leaves, fruits and seeds. The plants in the swamp populations have a shrub-like growth form, branching out very close to the ground, while those in forest populations have a tall, straight architecture. The swamp habitat is characterized by an open canopy, seasonal inundation, and thin or no soil. Forest sites are characterized by dense canopies, little or no inundation and moderate amounts of soil. Swampy regions occur in many parts of the Yucatan peninsula, which are inadequate for most traditional agricultural uses (corn, beans and squash). The chicozapote could become a valuable crop plant in these areas.

The chicozapote is a variable, outcrossing species, which has historically led to many subdivisions of the original species observed by Linnaeus. In Pennington's 1990 monograph on the Sapotaceae, 57 synonyms were listed for Manilkara zapota. Most of the synonyms were based on small variations in floral characters that were described by Pennington as being within the normal range of an outcrossing species. Based upon the observed morphological variation between the swamp and forest populations in the Yucatan peninsula, the populations may in fact be ecotypes of M. zapota, or they may be expressing phenotypic plasticity in response to the different environments. To date, no studies on the populations genetics of Manilkara zapota have been published, and I therefore attempted to learn more about this important plant by examining genetic variation.

The specific question addressed in this study is the following: Is there a genetic correlate to the extreme morphological differentiation observed between swamp and forest populations of M. zapota? If the populations found in the swamps are distinct ecotypes, they may be more suited for agriculture in these areas.

In order to test this question, genetic variation was analyzed using random amplified polymorphic DNA (RAPD) through the polymerase chain reaction (PCR). The use of RAPDs has many advantages for finding genetic variation (Williams et al. 1990), among which are the relatively low cost, no radioactivity, a low sample DNA requirement and the high frequency of polymorphic bands revealed. Another advantage of this technique in the remote locations of the study is that it is not necessary to carry cumbersome containers of dry ice or liquid nitrogen for sample tissue preservation. It was possible to desiccate the samples using silica gel, which preserved the DNA for several weeks, long enough to return the samples to the laboratory. Several researchers have used RAPDs to search for intraspecific genetic variation. This technique has been used to identify closely related cultivars of alfalfa (Yu and Pauls 1993) and oilseed rape (Lee et al. 1996). It has also been used to discover relationships among cultivars of cacao (Wilde et al. 1992) and Alstroemeria (Dubouzet et al. 1997). Schierenbeck et al. (1997) used RAPDs to characterize diversity in tropical tree species along an elevation gradient, and Lashermes et al. (1996) investigated diversity with RAPDs in cultivated and wild coffee. To quickly evaluate genetic variation among populations, the technique of bulked segregant analysis was used, as described by Michelmore et al. (1991). In this technique, DNA from each individual in a population is combined into a single sample (bulk). Primers can then rapidly be screened for variation without screening each individual for every primer.