The analysis of paternity and maternity in the marine hydrozoan Hydractinia symbiolongicarpus using randomly amplified polymorphic DNA (RAPD) markers


  • This work represents the initial stages of an analysis of population structure and the determinants of gender-specific variation in reproductive success in the clonal hydrozoan Hydractinia. Rick Grosberg and Don Levitan share interests in the evolution of life-history traits, sexual allocation strategies, clonal demography, and self/nonself recognition systems. Don Levitan is currently a postdoctoral fellow in the Cnidarian Reproductive Analysis Project in Rick Grosberg's laboratory at the University of California, Davis.

E-mail:; tel. (916)752–2937, fax (916) 752–1449


For organisms in which direct observation of mating and subsequent dispersal of offspring and relatives is impossible, patterns of reproductive success and genealogical relationship can only be established using genetic markers. The ideal genetic assay would (1) employ highly polymorphic genetic markers for distinguishing among individuals; (2) use little tissue for analysing early life-history stages; and (3) require minimal investment in time and money for population level studies. From this perspective, DNA polymorphisms revealed by PCR amplification using random ten-base primers [Randomly Amplified Polymorphic DNA (PCR-RAPD) or Arbitrarily Primed DNA (AP-PCR)] have great potential. However, the evidence that RAPD/AP markers are both heritable and can be repeatably amplified remains controversial. This study characterizes patterns of inheritance and polymorphism of RAPD markers in the free-spawning, colonial marine hydrozoan Hydractinia symbiolongicarpus. In all cases, the amplification products were identical among extractions from the same clone. Of 56 primers screened, 13 had sufficient polymorphism and scoreability for an analysis of parentage and higher-order genetic relationships in three matings. These primers generated 156 unique amplification products (putative loci), of which 133 were polymorphic. All but four of these loci were inherited as dominant mendelian markers. Our study suggests that the presence of a marker represents a single allele at a locus; however, what appear to be single null alleles may actually comprise several segregating alleles. When the identity of neither parent was known a priori, inclusion (unique markers present in offspring and only one of the potential parents) proved to be more efficient than exclusion for assigning offspring to parents. The most powerful approach, however, was cluster analysis of all presence/absence information for the marker bands. Clustering avoided the pitfalls caused by the appearance of occasional nonparental bands, and constructed a hierarchical framework that correctly reflected all genealogical relationships.