Radiolaria Polycystina
Morphology and classification systems

Based on his previous monograph of 1862, and especially on the extensive collections of HMS Challenger, Haeckel (1887) produced the first comprehensive system of radiolarian classification, encompassing over 3,000 species, 2,400 of which were new to science. Although Haeckel's work is still a compulsory reference guide for anyone attempting to deal with the identification of these organisms, it has for some time been evident that it does not satisfactorily represent natural relationships. Indeed Haeckel’s groupings are only based on morphological similarities without the support of continuity in the fossil record, rather than on demonstrable evolutionary sequences. In addition, the rigidity of his diagnoses, based chiefly on strict geometric considerations (classification of Sphaeroidea), ignores the ample intraspecific variability of the polycystines. As a result, many of his described "species" are but slightly different morphotypes or even developmental stages of the same organism (see classification of Sphaeroidea; stages of growth in skeletons 1; stages of growth in skeletons 2;stages of growth in skeletons 3).

In spite of these shortcomings and the time elapsed, advances in the development of a better classification system have been very limited. Efforts to depart from and improve upon the classification schemes inherited from earlier workers have mainly followed two different approaches: cytological techniques and evolutionary studies.

Hollande and Enjumet (1960), Cachon and Cachon (1972), Petrushevskaya et al. (1976), Petrushevskaya (1981) proposed revisions which use not only the skeleton (as most other classifications), but also cytoplasmic features, in particular the "nucleoaxopodial complex" (sensu Petrushevskaya, 1981). Although these schemes are probably sounder in biological terms, their applicability to fossil and subfossil materials lacking the protoplasm is problematic, which is one of the reasons for their very limited acceptance among radiolarian workers.

Analyses of evolutionary lineages in geological sequences were somewhat more succesful than cytological techniques in defining characters applicable to classification. Based on evolutionary evidence, Riedel and Sanfilippo (1986) produced an interesting critical review of the most important skeletal traits used by Haeckel. They concluded that some of them (e.g., number of segments, number of supplementary concentric spheres, number of feet, number of rays and of equatorial spines in discoidal Spumellaria, presence and nature of thoracic wings) have little or no suprageneric value. In contrast, several others (especially cephalic structure, but also pore arrangement, shell terminations in Nassellaria, etc.), traditionally considered as of minor value, are conservative through time, reveal evolutionary lineages and, therefore, are relevant for higher-rank divisions.

Riedel (1967b, 1971), Petrushevskaya (1965, 1971a), Goll (1968, 1969), Sanfilippo and Riedel (1970), Zhamoida and Kozlova (1971), Foreman (1973), Dumitrica (1988, 1989), De Wever et al. (2001) based on skeletal features alone worked out alternative classifications, either for the entire order or for selected polycystine groups. Of these, Riedel's (1967b, 1971) suprageneric system has become the most widely accepted for extant and Cenozoic radiolarians, and is the one adopted herewith (with slight modifications; see also Kling, 1978; Boltovskoy, 1981e; Anderson et al., 1996, 2000). It should be stressed, however, that this system does not overcome many of the above-mentioned problems, and is therefore a compromise provisional classification. Several of the family-level definitions, especially in the Spumellaria, are rather vague and generally used as a lumping black box for the many forms with complex morphologies and poorly understood relationships (e.g., Family Litheliidae, Family Pyloniidae, Family Tholoniidae).

Specific identification of the polycystines is a time-consuming and frustrating task. With the exception of the few abundant and widespread species on whose names there is fairly good agreement, binomial nomenclature alone very often fails to pinpoint unequivocally a given morphotype because different names are applied to the same species and, conversely, identical organisms are reported under different specific and even generic names (see Boltovskoy and Jankilevich, 1985). Because a very substantial proportion of the original species descriptions were published in old and often hard to get monographs, some authors find it faster and easier to create a "new species" for the unusual-looking skeleton in the slide, than to comb the dusty books in search of an adequate, already established name. Ecologically, paleoecologically and stratigraphically-oriented studies often underestimate the importance of a stable and consistent naming system; the lack of species illustrations in these reports allows the wrong designations to go undetected. This not only hinders buildup of useful information, but also significantly degrades the overall quality of radiolarian-based data for other applications. Recent literature has abundant examples of this bias, which introduces even more chaos into the already anarchic situation inherited from turn of the century works. Indeed, this may be a major reason for the waning use of radiolarians in stratigraphic and paleoecologic work.

The illustrated glossary of most commonly used terms for the description of polycystine skeletons is chiefly based on the listing compiled by Petrushevskaya (1981). Note: all these terms are included in the Glossary module; the explanatory text is preceded by "[Radiolaria Polycystina]".