Many Solanum fruits contain high levels of glycoalkaloids (GA's). While most species have been reported to lose nearly all GA's with fruit ripening, some plants such as S. carolinense (Horsenettle) have been shown to retain high levels. Because high levels of compounds potentially toxic to vertebrates seems paradoxical in ripe fruits, we sought to further explore GA patterns in temperate and tropical Solanum species. We used two methods to examine fruit chemistry: 1) a colorimetric "total GA" assay that measured total aglycone (solasodine), and 2) reversed-phase HPLC to quantify the principal GA's. Most fruits were collected from wild plants, therefore patterns of variation could be due both to genetic and environmental sources. But for S. carolinense, we also examined variation among and within 10 clones grown in a common garden. We studied unripe fruits (those that had attained full size, but were still green) and ripe fruits (those that had completely changed color and had started to soften.) Our results supported the suspected pattern of decreases in glycoalkaloids with ripening in many species and little or no decrease in S. carolinense fruits. Intraspecific variation was 10-20 % (C.V.) in S. carolinense clones, depending on method. When ripe fruits were sorted according to seed dispersal syndrome, we found that bird fruits contained very little or no GA (2-8 mg/g) and small mammal fruits had the highest levels (32-44 mg/g total GA). Bat fruits were variable, ranging from 2-27 mg/g total GA. We discuss our results in relation to patterns of seed dispersal and the degree to which past and current selection pressures might influence patterns of fruit-pulp chemistry.

Key words: frugivory, fruit secondary chemistry, glycoalkaloids, seed dispersal, Solanum, steroids