The worm endosymbionts in tabulate corals from the Silurian of Podolia , Ukraine

Two endosymbionts, Chaetosalpinx sibiriensis and Coralloconchus bragensis, occur in Silurian tabulate corals of Podolia. The endosymbiotic worms responsible for C. sibiriensis bioclaustrations in tabulates are found only in certain species: Paleofavosites cf. collatatus, Heliolites sp. A, Heliolites sp. B, Heliolites sp. C, Favosites gothlandicus, Favosites sp. A. One to six C. sibiriensis-infested tabulate species are known from Late Homerian to Ludfordian, in the reef-related community. Chaetosalpinx sibiriensis preferred certain tabulate species over the others, but showed no preference for the favositid or heliolitid type of morphology.

The Silurian rocks (Wenlock-Pridoli) of Podolia (Ukraine) are exposed in an approximately 80 km wide area along the Dniester River and its tributaries (Fig. 1).The Silurian deposits were formed in variable conditions from the normal-marine to lagoon facies (Tsegelnjuk et al. 1983).The massive coral-stromatoporoid-algal bioherms (Grytsenko 2007) from Lower to Upper Ludlow of Podolia are characteristic of a shallow shelf environment (Fig. 2).
The aim of this paper is to test the following hypotheses: (1) the endosymbiotic worms occurred only in certain host species; (2) the infestation rates are hostspecific; (3) endosymbionts preferred a certain type of tabulate morphology (heliolitid versus favositid); (4) the number of infested tabulate species in the coral reef and reef-related community changes over time; (5) the infestation rate of coral species changes with time.

Two
The infestation pattern of tabulate corals by C. sibiriensis is not random.The worms responsible for C. sibiriensis bioclaustrations were capable of infesting certain tabulate species: Paleofavosites cf.collatatus, Heliolites sp.A, Heliolites sp.B, Heliolites sp.C, Favosites gothlandicus, Favosites sp.A (6 of the 19 species studied).One of these species was infested both in the Late Homerian and Gorstian, one in the Gorstian as well as in the Ludfordian, one in the Ludfordian, and three infestations occurred in the Gorstian (Tables 1 and 2).The infestation rates of P. cf. collatatus changed only moderately over time (Table 3).
Chaetosalpinx sibiriensis preferred the tabulates P. cf. collatatus (71% of the specimens infested) over Heliolites sp.B (25% infested) and Favosites gothlandicus (11% infested) in the Gorstian, however, no preference was observed concerning the favositid (infestation rates from 11% to 71%) or heliolitid (infestation rates from 25% to 42%) type of tabulate morphology.The multiple peaks on the histogram of the C. sibiriensis diameter (Fig. 5) in most infested P. cf. collatatus indicate that worms responsible for the bioclaustration could have had more than one distinct growth type.

DISCUSSION
The number of infested tabulate species in the Silurian of Podolia increases from one in the Late Homerian to five in the Gorstian, and decreases to two in the Ludfordian (Table 2).However, these numbers are based on relatively small samples, while different results can be obtained from larger samples.
It is not possible to detect on the fossil material whether variations in the infestation rates in the Gorstian (Table 3) result from active larval selectivity by the worms or simply indicate their success of colonization.The environmental differences in worm infestation rates could be an alternative way to explain the species-specific percentage of infested specimens, if the infested corals lived in different environments.All studied tabulates were collected from shallow shelf facies and we believe that all our C. sibiriensis-infested corals were living in a similar environment.Sokol (1; 0) Zhvanets (2; 1) Ustje (2; 0) Tsviklevtsy (1; 0) Tsviklevtsy-186 (1; 0)

Favosites gothlandicus
Sokol-25 (10; 0) Grinchuk-27 (11; 0) Tsviklevtsy-186 (6; 1) Ustje (2; 1) Malinovetskaya sloboda (1; 0) Sokol (1; 0) Favosites sp.A Ustje (2; 1) Tsviklevtsy-186 (1; 1)   Chaetosalpinx has been interpreted as a tabulate parasite, considering its position between the corallites, perforation of the host's skeleton and soft tissue, modification of its phenotype and possible inhibition of its growth (Zapalski 2007).Thus, some of the uninfested tabulate species could have evolved active means of protection against infestation by C. sibiriensis.On the other hand, as we did not find any adverse effects of Chaetosalpinx on the neighbouring coral skeleton (Fig. 3), we suggest that the relationship was commensal.In Palaeozoic endosymbiosis it is possible that the particular tabulate coral taxa which tend to contain bioclaustrations may have been among the community's least aggressive or toxic corals (Tapanila 2005).Three peaks on the histogram of C. sibiriensis diameter in P. cf. collatatus (Fig. 5) could be explained by the presence of three distinct morphotypes.Thus, there may have been more than one species of endosymbiotic worms responsible for C. sibiriensis in P. cf. collatatus.We did not find any correlation between the growth form of the corals and the occurrence of Chaetosalpinx bioclaustrations.
In Palaeozoic tabulate corals bioclaustrations are preferentially found in common host taxa that span millions of years of geologic time, e.g.Favosites (Tapanila 2005).Our observations on the Silurian tabulates from Podolia support this opinion.We found the highest infestation rates in P. cf. collatatus, which is a common tabulate species in the Muksha Subformation (Mõtus & Grytsenko 2007) and ranges from the Late Homerian to the Gorstian (Table 1).Common and longranged species had sufficient abundance in space and a long exposure time for their potential symbionts, increasing so the probability of the development of an endosymbiotic relationship as compared to short-ranged and less common coral species.
In Baltica Chaetosalpinx has hitherto been reported from Favosites vicinalis (Stel 1976), Heliolites sp.(Stel 1976), Parafavosites germanica (Sokolov 1948), and Paleofavosites balticus (Klaamann 1958), and probably also from Favosites pseudoforbesi (Stel 1976) and Thecia swindereniana (Stel 1976).Our new data from Podolia show that Chaetosalpinx was infesting many more coral species in the Silurian of Baltica, and besides favositids, also heliolitids were common hosts.Chaetosalpinx is hitherto known from eight species of favositids (Tapanila 2005; our data), three species of heliolitids (our data), and one species of Thecia (Stel 1976).Thus, in the Silurian favositid corals could have had the highest number of Chaetosalpinx-tolerant species as compared to the rest of tabulates.However, among the Chaetosalpinx-tolerant corals no preference of favositids over heliolitids is observed as both the highest and lowest infestation rates characterize favositids (Table 3).The scanty data on the occurrences of Chaetosalpinx outside Baltica do not allow analysing the palaeobiogeographic distribution patterns of the genus in the Silurian.

CONCLUSIONS
In Palaeozoic tabulate corals bioclaustrations are mainly found in common host taxa spanning millions of years of geologic time, e.g.Favosites (Tapanila 2005).We established the highest infestation rates in P. cf. collatatus, possibly because this species is a common tabulate in the Muksha Subformation.Among the Chaetosalpinx-tolerant corals, favositids show no preference over heliolitids.Both the highest and lowest infestation rates are associated with favositids.

Table 1 .
Stratigraphic and locality information of Chaetosalpinx sibiriensis-infested tabulate species.The first number in parentheses shows the number of the studied samples, the second number in parentheses -the number of infested tabulate specimens

Table 3 .
Infestation rates of the tabulate coral species by Chaetosalpinx sibiriensis in the Silurian of Podolia