crystallization of siliceous gels ( Lebedev 1967), but the 

 solutions forming such gels are dense viscous liquids of much 

 greater silica concentration than modern hydrothermal fluids. As 

 well, in the present example the formation of filamentous silica 

 is directly linked to the morphology of the underlying support of 

 iron-oxide filaments upon which the silica precipitates. The 

 formation of dendrites, filaments of iron-oxide, also requires 

 the presence of a solid support. Thus, it is difficult to 

 explain the upward growth of filamentous iron/silica deposits 

 without invoking the prior existence of a filamentous structure 

 that would have acted as a support for mineral precipitation. The 

 similarity of these filaments in size and shape to those produced 

 by bacteria has led us to consider how the deposits could have 

 formed within a filamentous structure provided by micro- 

 organisms . 



Filamentous micro-organisms are very abundant around 

 hydrothermal vents, as this morphology is common in groups of 

 sulfur and metal-oxidizing bacteria ( Jannasch and Wirsen 1981). 

 Branching filaments are best known among the iron and manganese 

 oxidizing taxa. It is quite conceivable that mat-like 

 aggregations of iron-accumulating bacteria could have existed at 

 the sites where the described deposits were formed. Growth of 

 microbial mats around vent openings would result in the 

 diffusion of hydrothermal fluid through the filamentous 

 structure. A combination of bio-catalysis and autocatalysis 

 similar to that proposed above for vestimentif eran tubes could 

 lead to iron accumulation on the filaments. Mineralogical 

 zonation observed in the deposits could result from differential 

 mixing of vent fluid and ambient seawater with the mat structure 

 (Juniper and Fouquet 1988). Similar physico-chemical gradients 

 have been observed in seafloor mats of the filamentous sulfur 

 bacteria Beggiatoa (Nelson et al. 1986; Moller et al. 1985). 



The precipitation of amorphous silica from hydrothermal 

 fluid requires more than a structural support. The inorganic 

 formation of Opal A is rare because it requires a medium 

 saturated in silica (Bertine and Keen 1974, Kastner 1979, Wollast 

 1974). While hydrothermal fluids are rich in silica, cooling by 

 mixing with seawater does not normally result in silica 

 precipitation because the resulting dilution prevents saturation 

 (Wollast 1974). Substantial conductive cooling is necessary to 

 precipitate silica from hydrothermal fluids. The 

 microenvironment provided by a porous filamentous structure, 

 partially filled by iron-oxide precipitation could sufficiently 

 limit mixing and allow enough conductive cooling to permit 

 silica saturation and Opal A precipitation to occur. A similar 

 process is likely responsible for in-filling of porous sulfide 

 chimneys by amorphous silica. The fact that silica deposition 

 occurs consistently after iron suggests that it is the slower 

 process. 



114 



