no RIVERSIDE LETTERS XIV 



mingled with the general roughened surface. 

 These are always long strip-shaped pieces, 

 flowing out with the stream, which keep con- 

 tinually altering their curving forms. I can- 

 not account for these smooth places except by 

 supposing that they are due to some kind of 

 oil on the water ; my reason for this supposi- 

 tion is that I have noticed they very fre- 

 quently have as their starting point some little 

 patch of rushes or weeds ; these rushes, I 

 believe, as they get stirred by the wind set 

 free an oil from the mud around their roots, 

 which rises to the surface and flows off down 

 the stream. The fact that tiny streaks of 

 smooth water can be seen below even solitary 

 rushes lends strength to my theory. That 

 the rushes cause the smooth places by the 

 shelter they afford from the wind cannot be 

 entertained, as these streaks of smooth water 

 flow always from the rushes down stream 

 even when the wind is blowing directly up. 

 These markings on the river surface, during 

 rain, have, as far as I remember, never been 



