In contrast, changes in total CO, dur- ing storage fell within the analytical error of 2.5 percent. The curves (Fig. 1) were calculated for 25°C from the data of Lagerstrom (//), as verified by Ingri (72), for equilibrium constants of silica species in solution. According to Fig. 1, our data can be divided into four groups: (a) concentrated, intracrustral high-pH brines from Lake Magadi and Alkali Valley; (5) concentrated brines from open pools in Alkali Valley and Lake Magadi, from the brine pool north of Abert Lake, and from the closed sag pond of Deep Springs Valley; (c) brines of intermediate concentration, interstitial to saline muds from Albert and Magadi takes; and (d) Abert Lake, plus saline springs and interstitial fluids from adjacent mud flats, inflow brines from springs of the Lake Magadi area, and brines from the area of Deep Springs Lake. No systematic correlation exists between silica content and temperature at the time of collection of each brine, although some of the scatter (Fig. |) may be due to temperature differences. Groups (a) and (d) are under- saturated with respect to amorphous silica, group (c) is of intermediate pH and near saturation, while group (5) is close to supersaturation or substan- tially supersaturated. The only other nongeothermal water reliably reported to have a very high pH andsilica content, from Aqua de Ney Springs, Cali- fornia (/3), plots very near the satura- tion curve for amorphous silica. The correlation between pH and SiO. for each group can be explained by their special history. We believe that the enrichment in silica in the brines is primarily due to evaporative concentration. Points of group (a) show a high pH because of the depletion in HCO,-, caused by trona precipitation; the pH has increased more rapidly than evaporative concentration of SiO.; exchange The brines in contact with saline muds, containing abundant silicates [group (c)], plot near saturation with respect to amorphous silica. Higher SiO., values are from samples of bottom sediment near the water-mud interface; lower values, from interstitial waters deeper in the sediment. This fact suggests that initial equilibrium is established with respect to an amorphous surface layer on silicates (14); subsequent recrystallization lowers SiO, content. The compositions of the springs of Lake Magadi [group (d)] are governed by their underground history. A reverse trend appears to exist between pH and SiO. and probably reflects adsorption of with the atmosphere is apparently too silica on freshly precipitated sesquioxides (/5). The other points of group (¢d) are for waters of Abert or Deep The open pools of group () are either undersaturated with respect to have been concentrated from waters having distinctly lower SiO, content. slow to checkthis trend. sodium carbonate minerals or saturated with natron (NA.CO,:10H.,O). Evaporative concentration of SiO., is apparently more rapid than nucleation and precipitation of silica. Springs Lake: they are either dilute or If silica is indeed concentrated by evaporation, this concentration should be reflected in a correlation between SiO., and sodium, the predominant cation of these brines (Table 1). As Fig. Table 1. Representative alkaline brines from Alkali Valley (Alkali Lake, AIL) and Abert Lake (AL) basin, Oregon, Deep Springs Lake (DSL) and sag pond, Calif. (Deep Springs Valley, DSV), and the Lake Magadi (LM) area, Kenya. Included are brines with relatively high contents of CO:, even though chloride or sulfate may in fact dominate the anions. Little Magadi Lake, Kenya, LML. Sample pH Contents (ppm) Collece W i tion temp.. time (°C) Field Labora- 120°03’°W ,42°S7‘N 8~6-63 25 9.97 Surface brine from pothole near log road on SW side, AIL playa 120°03'W ,42°S7°N 8-6-63 40 At gage, AL 120° 117W,42°36’N 7-22-64 23 9.75 At mouth of Chewaucan River, surface, AL 120°15’°W 42°3 VN 7-25-64 30 9.82 Brine from within very porous salt crust, central DSL 118°02°W,37°17’N 8-18-63 41 8.9 Closed sag pond, DSV 118° L.5°W,37°17VN 8-15-61 33 Interst, brine from bottom mud, $ end, near W shore, AL 120° 15’°W ,42°32’N 8-6-63 22 9.05 Interst, brine from bottom mud, SW end, AL 120°15'W ,42°33’N 8-6-63 22 8.65 148 Interest, brine from bottom mud, near mouth of Chewaucan River, AL 120° 15’°W,42°3 WN @-8-63 36°16°E,01°44’S 6-66 Saturated brine, LM 36° 15’E,01°54’S 6-66 Saturated brine, LM 36° 17°E,01°S0’S 6-66 Description Brine from pit next to pothole near log road on SW side, AIL plava Hot spring, LML 1312 . Source coordinates 9.90 22 81 (COs) fate (SOx) Chioride 1 (C1) Dis(ane) trace 99,500 20,200 19,800 237,000 8080 trace 133,000 25,000 27,900 318,000 15,200 604. 3840 6920 792 12,700 39,300 2360 112 758 972 135 2190 6220 78 108.000 22,000 4990 20,700 56,200 120,000 332,000 §24 119,006 10,900 3630 66,160 90,900 41,000 330,000 211 19,500 1360 10,200 6370 949 16,900 50,300 20.800 1350 14,600 4780 1270 16,800 §2,300 1200 4280 35180 1860 13,800 42,700 12,606 239 45,600 3540 147 5950 30,200 tory Silica 1). (SiO2) Sodium otas ; bonate 10.0 590 91,400 5090 10.4 1270) 121,000 9.65 128 9.85 48 10.0 9,25 162 16,100 sum (HCO;) cate“ solved id 9.05 9.05 90 34 11.06 10.25 1055 132,000 2280 trace 106,060 219 84,400 324,000 36 10.11 o.5 383 110.000 1530 trace 95,600 97 50,700 247,000 SCIENCE, VOL, 158