The data of Anderson and Jones (1976) show that the macrophyte zone in shallow lakes larger than about 500 ha normally occupies < 25% of the lake area. Anderson (1978) was able to predict, with a combination of four maximum tolerance values, the presence or absence of sago in 136 of 140 sample sites in a large Manitoba wetland. No sites with a maximum fetch of 2,300 m, north fetch > 600 m, northwest fetch > 500 m, and water depth > 120 cm supported sago. Lake Ichkeul in Tunisia (90 km2) supports about 20 km2 of sago (Skinner and Smart 1984), but the plant is absent or sparse in the central area and along several long stretches of shoreline, possibly because of excessive fetch. Kantrud (1986a) found a similar absence of sago in the central area of maximum fetch in a large (1,132 ha), shallow North Dakota wetland. Ongoing studies of Butler and Hanson (1985, 1986, 1988, unpublished) show that in a large (1,598 ha), shallow (mostly < 4 m) Minnesota lake, concentrations of suspended solids are much greater (40-100 ppm) in central areas of greater fetch than in deeper, more sheltered sites (10-50 ppm). They used the formulas of Carper and Bachman (1984) to show that even moderate winds could expose large areas of the bottom to wind energy sufficient to cause sediment resuspension and increased turbidity. On large lakes, littoral zones can be barren due to wave action, with sago and other macrophytes restricted to river inlets (Sheffer and Robinson 1939). In other large lakes, wave action can restrict sago beds to deeper (> 3-4 m) sites where plants grow in sparse clumps (Hill et al. 1975). Pip (1987) observed no significant differences in frequency of sago in wetlands larger or smaller than 10 ha in area.