[ref. ID; 7307 (Eigner & Foissner, 1994)]

Family Amphisiellidar Jankowski, 1979
Euhypotrichina with an amphisiellid median cirral row (ACR) of which the anterior segment is formed by cirri of the rightmost ventral anlage and posterior segment by cirri of the second ventral anlage from right. A middle segment may be formed by neighboring anlagen. This diagnosis does not clearly exclude the Amphisiellidae from the Discocephalina as diagnosed by Wicklow (1982). His diagnosis of these obligatorily psammolittoral forms is rather wide and was sufficient for the data available at that time. To separate these families unequivocally, a refined (narrower) diagnosis of the Discocephalina is necessary: Euhypotrichina with a discocephalid median cirral row of which each cirrus originates from a different ventral anlage. Our diagnoses elucidate the different modes of forming the median cirral row in the Amphisiellidae and the Discocephalidae, respectively. To strengthen this difference, the use of different terms, viz. "amphisiellid median cirral row" and "discocephalid median cirral row", seems appropriate. Wicklow (1982) suggested the term "midfrontal cirral row" for the Discocephalidae. Although this term is well chosen, there may be problems with finding appropriate terms for the other modes of forming a median cirral row, e.g. for Orthoamphisiella. We thus suggest, to put simply the name of the main taxon (family) in front of the "median cirral row".

The amphisiellid median cirral row (ACR)
Migration of cirri takes an essential part in forming the final cirral pattern in many hypotrichs. The migration is often most distinct in the rightmost anlage and results in so-called frontoterminal cirri (urostylids and oxytrichids) or, as in amphisiellids, in a median cirral row composed of two or more anlagen segments. However, a median cirral row can be formd by at least four basic, non-homologous processes: It may develop in the proter and opisthe from single anlage within the parental median row. This mode occurs in Orthoamphisiella spp. (Eigner and Foissner 1993). Hemberger (1982 dissertation) depicts similar development in Uroleptoides atypicus, A. terricola and Trachelochaeta gonostomoida; these taxa possibly constitute a new family. Furthermore, the median cirral row may develop from a single anlage each in proter and opisthe. This mode occurs in Cladotricha koltzowii (Borror and Evans 1979) and in Engelmanniella mobilis (Wirnsberger-Aescht et al. 1989). Alternatively, the median cirral row of amphisiellids develops from two or more rightmost anlagen in both proter and opsithe. This mode occurs, e.g. in Hemiamphisiella terricola, Amphisiellides illuvialis and Paramphisiella caudata. All species and genera mentioned above form their median row by complete anlagen or at least large portions of them. In contrast, the median cirral row of the Discocephalidae, e.g. Psammocephalus faurei (Wicklow 1982), is formed by individual a cirri each originating from a different anlage. Paraurostyla weissei also possesses a cirral row that develops from two rightmost anlagen (Wirnsberger et al. 1985). However, several ventral rows, which evolve from independent primordia, occur left of this composite row. This produces a cirral pattern which is rather different from that of the amphisiellids. Nevertheless, a certain relationship between Paraurostyla, Amphisiella and Gastrostyla is apparent and supported by a phylogenetic approach using other apomorphies (Wirnsberger et al. 1986). In Onychodromus quadricornutus the rightmost anlage migrates only slightly anteriorly; a composite cirral row is not formed (Foissner et al. 1987). Moreover, as in P. weissei, several ventral rows occur in this species, which also separate it from the amphisiellids. Seven types (genera) of forming the ACR, the postperistomial cirrus, and the transverse cirri can be distinguished in amphisiellid hypotrichs.
The first type, found in Amphisiella spp., builds the ACR from cirri of the two rightmost anlagen. The anterior segment is formed by anteriorly migrated cirri of the rightmost anlage 6; the posterior segment is formed by cirri from anlage 5. Transverse cirri are formed either by posteriorly migrated cirri from these two rightmost anlagen or from more than two anlagen.
The second type, found in Amphisiellides illuvialis, builds the ACR as Amphisiella, but several cirri of the rightmost anlage remain right of ACR. Several posterior cirri of the rightmost anlage 5 migrate posteriorly to form a short, longitudinally arranged transverse cirral row. The second anlage from right (Berger and Foissner 1989) does not contribute to the transverse cirri.
The third type, found in Paramphisiella caudata, builds the ACR and the transverse cirri as Amphisiellides. However, no cirri remain right of the ACR.
The fourth type, found in Paragastrostyla lanceolata, builds the a ACR from three or four rightmost anlagen. The anterior segment of the ACR is formed by all anteriorly migrated cirri of the rightmost anlage 7; thus, no transverse cirri are formed in this genus. A middle segment is formed in this genus. A middle segment is formed by posteriorly migrated cirri of anlage 5; possibly anlage 4 is also involved in forming the ACR. The posterior segment is formed by all cirri of the second anlage from right (Borror and Evans 1979).
The fifth type, found in Gastrostyla steinii, builds the ACR from the three rightmost anlagen as Paragastrostyla. However, only the first (anteriormost) cirrus of the third anlage from right (Berger and Foissner 1989) forms the middle segment of the ACR; the second cirrus of this anlage (Berger and Foissner 1989) migrates further posteriad to form the postperistomial cirrus. Transverse cirri are formed by posteriorly migrated cirri of the rightmost and several neighboring anlagen. Hemberger (1982 dissertation) described the ventral row (ACR) of G. steinii a being composed of only two anlagen (segments). We assume, however, after a careful examination of his drawings, the ACR to be usually composed of a addition middle segment, viz. the anteriormost cirrus of anlage 4. However, alignment of this cirrus with the ACR may be imperfect in some specimens.
The sixth type, found in Hemiamphisiella terricola, forms the ACR and the postperistomial cirrus by processes as described for Gastrostyla, whereas the transverse cirri are formed as in Amphisiellides and Paramphisiella.
The seven type, found in Pseudouroleptus caudatus, forms the ACR and the postperistomial cirrus as the former two types. However, the rightmost anlage 6 produces a great number of cirri, of which most remains right of the ACR. These cirri form a distinct second ventral row which is obviously homologous to the transverse cirri and the cirri right of the ACR found in Amphisiellides.
These data show that the median cirral row of amphisiellid hypotrichs is built by at least two rightmost anlagen which arrange one upon the other during cytokinesis and in post-dividers. We consider this unique median cirral row formation as an apomorphy defining the family Amphisiellidae. The alignment of the anlagen is frequently imperfect, i.e. the ACR usually appears more or less distictly interrupted in interphase specimens. This irregularity is highly characteristic for amphisiellid hypotrichs.

Genus separation im amphisiellid ciliates and improved diagnoses
In spite of the homologous development of the ACR, morphogenetic differences in amphisiellid ciliates are distinct enough to maintain the genera proposed in the literature. We did not find any significant variation in the main characters traditionaly used for genus distinction in hypotrich ciliates, e.g. presence/absence of transverse and/or caudal cirri. However, when analyzing a great number of specimens of a certain species or genus, there are usually some individuals which deviate, e.g. with a missing postperistomial cirrus or lacking transverse cirri. Based on the ontogenetic data discussed above the Amphisiellidae Jankowski, 1979 and the genera contained therein can be defined more properly. We recognize seven genera, two of which were originally assigned to the Oxytrichidae.

Phylogeny of amphisiellid ciliates
The seven genera assigned by us to the Amphisiellidae have a rather similar interphase morphology and construct the median cirral row in a complex and very likely homologous manner. We thus expected that it should be easy to elucidate their phylogeny with Hennig's cladistic method (Ax 1984, Hennig 1982). Unfortunately, all trails to find convincing synapomorphies for most generic groups failed, possibly because of the occurrence of many convergencies which make determination of other character states (apomorphies, plesiomorhies) extremely difficult. To mention only two main problems: the number of anlagen comprising the ACR is very likely correlated with the occurrence of a postperistomial cirrus formed by a complicated cirral migration. Unfortunately, these two characters do not completely match and thus one has to assume that the postperistomial cirrus has been secondarily lost in Paragastrostyla. Likewise, dorsomarginal rows, independent primordia, and development of transverse cirri from more than one anlage varied rather arbitrarily in the schemes examined. Considering the these problems, we even cannot exclude the possibility that our family character, viz. the specific mode of forming the median cirral row, evolved more than one time. Apparently, even very detailed interphase and ontogenetic data are insufficient to puzzle out the bewildering evolutionary paths in amphisiellid (and other!?) hypotrichs. Complementary methods, like molecular markers and transmission electron microscopy of dividing cells must be applied.