The observed bioerosion structures raise several fundamental problems concerning the trace-making process: the supposed identity of the organisms responsible, the distribution of traces and the ichnotaxonomy of these traces. Few observations have been made concerning the identity of bioeroders of foraminiferal test, which include invertebrates (Sliter 1971; Walker 1971; Mageau and Walker 1975; Hickman and Lipps 1983; Arnold et al. 1985; Gooday 1986; Langer et al. 1995) and other foraminiferans (Hallock and Talge 1994; Nielsen 2002). The morphological variation of these bioerosion traces (Nielsen and Nielsen 2001) and the new traces described herein suggests that a number of unknown organisms should be considered as tracemakers. The outline and regularity of the traces seem to exclude most marine bacteria, fungi and algae as likely trace producers of them (cf. Vogel et al. 1995; Glaub and Bundschuh 1997; Perry 1998).

If any relationship exists between size of the bioerosion traces and size of the trace-making organisms or their physiological features as suggested by several authors (e.g., Taylor et al. 1980; Gibson and Watson 1989; Kowalewski 1993; Nebelsick and Kowalewski 1999), then many of the suggested organisms are simply too large to be likely candidates. The juveniles of these organisms may have the right size of body or physiological feature to be candidates but owing to the morphology of the bioerosion traces, these can probably be disregarded as well. The physiological feature used to produce the plug traces, Fossichnus solus, would need a morphology much like a hollow tube to produce such traces. Fossichnus solus may occur in groups as multiple specimens in the same foraminiferal test. As the trace apparently is an early stage of Oichnus simplex, which penetrates the test, the target may have been the cytoplasm of the foraminiferal test. However, most specimens of F. solus have not been transformed into O. simplex. As the way of penetration appears to have been inefficient, predatory behaviour is less likely. Instead, F. solus is more likely to be interpreted as an attachment scar formed by an unknown tracemaker.

As Oichnus aff. O. asperus and Oichnus aff. O. paraboloides provided the tracemakers with access to the interior of the foraminiferal tests, these trace fossils are best interpreted as predatory traces and may be categorized as praedichnia (Ekdale 1985). However, the physiological actions by which the trace fossils were formed remain unknown.

An organism having an attachment pedicle or pedicle-like feature could produce a trace like the sieve-shaped boring. However, we would like to suggest a second possibility, even though this is highly speculative. As can be seen in Figure 11, several aggregates of sediment grains are found attached to the aboral end of the foraminiferal test surfaces containing the sieve-shaped boring. These could be organic structures made by an arenaceous organism. Owing to the treatment of the specimen and because these structures have not been observed on any other foraminiferal test in the samples, it is unlikely that the grains represent non-organic structures. If this argument is correct, then a likely candidate as tracemaker could be this arenaceous organism. Most arenaceous organisms are found in the unicellular marine order of Foraminiferida. It has long been known that the foraminiferans have the ability to cause bioerosion for various reasons (Todd 1965; Banner 1971; Baumfalk et al. 1982; Plewes et al. 1993; Cedhagen 1994; Hallock and Talge 1994; Vénec-Peyré 1988, 1996; Collen 1998). Most of these species are calcareous but many arenaceous species live on the test of other marine species and at least one is known to be a predator (Cedhagen, personal commun., 1999). This is of course speculation since we do not know whether the aggregates are indeed a species of Foraminifera. Further, it is not known for certain whether any connection exits between the sieve structure and the sediment aggregates. This conundrum could be resolved by careful removal of the aggregates from the test surfaces. However, as this process would probably destroy the specimen, we have been unwilling to do so.

The occurrence of repeated solitary borings in different chambers, as seen in Figure 12, on the test of smaller planktonic foraminiferans, seems to confirm that unknown predators prey on these while living. The sequences of circular borings have so far only been observed in planktonic species smaller than 350 µm. Since these only occur in smaller specimens of planktonic foraminiferans it seems to exclude scavenging and favours predation as a likely explanation of these. The boring shown in Figure 13 may be interpreted as either a case of abandoned predation, a repair or an attachment structure. As no invertebrates were observed in the sample, this favours the two first interpretations. No etching structures are observed, which could imply that the trace is best interpreted as a repair structure.

Trace-fossil assemblages have been successfully defined for both soft and hard substrates to show the relationship between tracemakers and physical characters of the palaeoenviroments (e.g., Seilacher 1967; Vogel et al. 1995; Bromley 1996; Glaub and Bundschuh 1997). Studies of recent bioerosion of the foraminiferal test have shown that such a relationship can be demonstrated (Nielsen et al. 2002). However, until now, we have been unable to transfer these results to palaeoenvironments.

This is the first time that the relationship can be documented in a study of bioerosion of the foraminiferal test. The ichnospecies Fossichnus solus has only been observed in five out of 200 investigated samples. The age of the E-series is probably Holocene (see Nielsen and Nielsen 2001), whereas the samples from Rhodes, P 56 and 194, are Early to Middle Pleistocene. A preliminary interpretation of the foraminiferal fauna found in the two Pleistocene samples is that P 56 represents a very diverse, fully oxygenated fauna, deposited in water depth of more than 300 m. The foraminiferal fauna from sample P 194 represents a poorly diverse assemblage probably deposited in water depth somewhat deeper than the P 56 fauna and during conditions when the bottom-water mass was somewhat deficient in oxygen. The occurrence of several species of temperate and subtropical planktonic foraminiferans seems to indicate that temperature and salinity would have been much like the present day conditions of deeper levels in the Aegean Sea.

The three samples of the E-series were obtained as box-samples from a water depth of 190 to 200 m. We have no information concerning the physical characteristics of the water mass, but the planktonic foraminiferans constitute a typical low diversity, cold-water fauna. The occurrence of Fossichnus solus in these two very different environments suggests that its distribution is controlled by water temperature and other parameters related to oceanography.