Phenotypic and genetic structure of interbreeding populations of the diatom Tabularia fasciculata (Bacillariophyta) IRENA KACZMARSKA 1 *, JAMES MICHAEL EHRMAN 2 ,MONICA BARROS JOYCE MONIZ 1 AND NIKOLAI DAVIDOVICH 3 1 Department of Biology, Mount Allison University, Sackville, NB E4L 1G7, Canada 2 Digital Microscopy Facility, Mount Allison University, Sackville, NB E4L 1G7, Canada 3 Karadag Natural Reserve of the National Academy of Sciences, Feodosiya, 98 188 Ukraine KACZMARSKA I., EHRMAN J.M., MONIZ M.B.J. AND DAVIDOVICH N. 2009. Phenotypic and genetic structure of interbreeding populations of the diatom Tabularia fasciculata (Bacillariophyta). Phycologia 48: 391–403. DOI: 10.2216/ 08-74.1 Diatom alpha-taxonomy heavily relies on the discontinuity in variation of species-specific structures in siliceous components of the cell wall. However, considerable intraspecific variability of valve morphology is to be expected because of the peculiar mode of diatom valve genesis during the tenure of the asexual part of the diatom life cycle. Here we evaluated variation in valve morphology among 66 Canadian and three Ukrainian clones of the diatom Tabularia fasciculata. Valve length, width, number of striae in 10 mm and number of rimoportulae per valve were enumerated for all clones. Forty-five Canadian and three Ukrainian clones were crossed and their sexual identity determined. Nuclear internal transcribed spacer 2 and 5.8S rDNA gene sequences were used to evaluate genetic variability of the clones. Our data show consistent phenotypic (at least two phenodemes ‘n’ and ‘w’) and genetic (two genodemes, ‘A’ and ‘B’) differentiation within one gamodeme (sexually compatible clones) in Canadian populations of T. fasciculata. The distribution of phenogenodemes is not continuous, even or random; a specific morphology strongly correlates a specific genotype, A and B, both members of the same gamodeme, with just a few exceptions. Specifically, width, striae density and rimoportulae number conform well (though not perfectly) to a specific genotype; only four mismatches were found among clones examined. Crimean members of the n phenodeme were morphologically and genetically very similar to those in Canada despite separation by thousands of kilometers. This is the first documented relationship between pheno- and genotype for an araphid species for which morphological, molecular and breeding data are also provided. KEY WORDS: Araphid diatom, 5.8S-ITS2-region, Reproductive compatibility, Phenodeme, Genodeme, Sympatric, Allopatric, Barcoding INTRODUCTION The current resurgence of interest in diatom alpha- taxonomy is largely a result of the development of new tools that expand species-specific diagnostic features beyond characters centered primarily on frustule morphol- ogy. Variability in valve morphology in particular has received much attention because of the rich microarchitec- ture of diatom valves on which to base species delineations. Numerous studies examined valve morphology in natural populations (e.g. Stoermer & Ladewski 1982; Hu ¨rliman & Straub 1991; Droop et al. 2000; Stachura-Suchoples 2006), under varying growth conditions (Syvertsen 1977; Wood et al. 1987; Cox 1995; Schmid 1997), culture-induced mor- phological abnormalities (Schmid 1980; Torgan et al. 2006) or phase of the life history (Cerino et al. 2005), mainly with the application of light microscopy (Mann 1999). Several more recent studies combined morphological and molecular characteristics, providing novel insights into the relationship between the species-specific range of morpho- logical variability (boundaries of morphospecies) and genetic variance in several coding and noncoding sequences from nuclear, mitochondrial and plastid-derived markers (Lundholm et al. 2002, 2003; Sarno et al. 2005; Evans et al. 2008). Numerous new species have been discovered using this approach (e.g., Mann et al. 2004; Lundholm et al. 2006; Sarno et al. 2005), where previously morphological plasticity was assumed. In other studies, the congruence between morphological, phylogenetic and biological species concepts were investigated (Behnke et al. 2004; Amato et al. 2007; Vanormelingen et al. 2007; Casteleyn et al. 2008; Kaczmarska et al. 2008), resulting at times in inconsistent findings regarding the goodness of fit between different species delineators. In some diatom species, different, sometimes very subtle morphologies, correlated well with different genotypes (semicryptic species; Mann et al. 2004; Amato et al. 2007; Casteleyn et al. 2008). In other species, both centrics and pennates, valve morphology was a poor predictor for the existence of reproductively isolated genotypes (cryptic species; Beszteri et al. 2005; Evans et al. 2008; Kaczmarska et al. 2008). With many of the studies discussed above leading to discovery of unanticipated diversity and the description of new species (both morpho- and phylogenetically defined), the range of intraspecific variability in biologically defined diatom species valve morphology remains poorly under- stood. Can the relationship between diatom genotypes and phenotypes be detected at the intraspecific level? Or do interbreeding diatom populations occur in a continuum of genetic and morphological variance? If there are detectable phenotypes and corresponding genotypes in diatoms, they * Corresponding author (iehrman@mta.ca). Phycologia (2009) Volume 48 (5), 391–403 Published 4 September 2009 391