TABLE 3.

Potential contribution of geosmin by geosmin-producing bacteria in water, intestinal mucous, digesta and on dorsal fin.

	Intestinal mucous	Flesh		Digesta	Fins	Water
Farm	Calculated geosmin production	Measured geosmin in flesh	Potential geosmin contribution by bacteria in intestinal mucous to content in flesh	Calculated geosmin production	Calculated geosmin production	Calculated geosmin production	Measured geosmin
	ng ⋅ intestine–1 ± SE	ng ⋅ kg–1 ± SE	% (average)	pg ⋅ g–1 ± SE	ng ⋅ cm–2 ± SE	ng ⋅ L–1 ± SD	ng ⋅ L–1
1	325.9	664.9	47.7	5.94.0	1.450.47	0.010.01	1.50.3
2	6410.9	21510.2	29.5	67.450.6	1.720.50	0.460.30	1.50.1
3	13633.0	28926.5	46.9	22.118.8	1.140.26	0.050.05	2.50.4
4	18642.9	43951.9	42.3	32.129.3	2.060.46	1.210.61	2.70.4
5	11426.1	40865.4	27.9	9.44.7	1.650.28	0.120.12	3.20.7
6	9419.3	35127.1	26.8	4.10.9	0.960.25	0	0

Assumptions were: (1) Each geosmin producing bacteria released 140 × 10^–18 g geosmin as observed for Streptomyces 2R; (2) the OTUs has been affiliated into the groups determined in Table 2; (3) for calculation of geosmin release from the entire intestine, surface area of the intestine in tilapia was assumed to be 2000 cm² (estimated from Frierson and Foltz, 1992). Weight of fish from each of the six farms (range 250 to 651 g) was used to convert potential fish-specific geosmin contribution by intestinal bacteria to weight-specific geosmin content in the flesh. Measured geosmin content in water and fish flesh refers to values shown in Figure 1. SE, standard error; SD, standard deviation.