microscopy (Fig. 9) the detailed structure of the mat is clearly 

 visible. The primary morphology is that of long, thin filaments 

 approximately 1-2 ym in diameter and up to 200-500 p m in length. 

 No filament branches were observed in any of our preparations. 

 Transmission electron microscopic analyses indicated that these 

 filaments were actually tubular in shape with relatively uniform 

 wall diameters of 0.2-0.3 M m (Fig. 10). Of approximately 1000 

 cross-sections that were analyzed, only 1-2% of the filaments 

 had recognizable cellular materials inside of the electron dense 

 walls (an example is presented in Fig. 10); the majority appeared 

 to be empty. At the present time we do not know whether this is 

 a result of poor preservation of these bacterial structures or 

 whether these initial results adequately describe the relative 

 proportion of "live" versus "fossil" filaments. In addition, 

 there was also a relatively large and morphologically diverse 

 assemblage of bacteria that were either attached to or otherwise 

 associated with the bacterial filaments. Because we made no 

 attempt to quantify these attached cells we are not certain 

 whether they comprise a significant portion of the total mat 

 biomass. 



We further investigated the nature of the filamentous matrix 

 to determine the bulk chemical and mineralogical composition of 

 these materials. These analyses included scanning electron 

 microscopy-energy dispersive x-ray fluorescence spectroscopy 

 (SEM-EDS) of individual filaments, x-ray diffraction, electron 

 microprobe analysis and bulk chemical analysis for phosphorus and 

 organic carbon. The latter analyses revealed that the mats were 

 only 1.8 + 0.07% organic carbon (by weight) indicating the 

 presence of a large inorganic component. The l^C/^C car b on 

 isotope ratio (expressed as S^C) was -24.3 o/oo (G. Rau, pers. 

 comm ) . 



The P content of the bacterial mat, as determined by 

 chemical analysis, was 2.4% by weight. This exceeds the total 

 organic carbon value and indicates that most of the P must be 

 inorganic rather than cell-associated (the average C:P [wt/wt] 

 ratio of bacteria under P sufficient growth conditions is 

 expected to be 30-50). From the results of our chemical leaching 

 experiments, we conclude that <2% of the P associated with the 

 bacterial mat is "CaC03-bound" (i.e., <2% is extracted in 

 acetate buffer, pH 5.0), <2% is "mineral-P" (difference between 

 oxalate buffer, pH 3.0 and 8 M HC1 extractions) and that the 

 majority (>95%) of the P must be termed "hydrogenous." Our 

 results support Berner's (1973) hypothesis regarding the 

 sorption or coprecipitation of P onto hydrothermal ferric 

 oxyhydroxides. These results are also consistent with the 

 experimental results of Froelich, Bender and Heath (1977) for 

 samples collected from the East Pacific Rise, and with the 

 stated implications for rapid P accumulation in hydrothermal 

 sediments. However, at Loihi Seamount it is possible that the 

 origin of the P that is accumulating in the metalliferous 



221 



