DEAD WILLOW BEND. 135 



mice tenanted most of these and fed largely upon the 

 delicate rootlets. I am inclined to believe that the bur- 

 rows are there because the trees happen to be growing 

 upon the banks of the creek, and the same would be the 

 case were any other tree planted or growing in their 

 stead. Still, I must admit that my examinations of 

 other localities where there were no willows did not 

 bear me out. 



On landing, I found the usual burrows, but all were 

 too large to be the work of meadow-mice. I endeav- 

 ored to probe them and determine what creatures occu- 

 pied them, but in this was altogether unsuccessful. No 

 mammals appeared, and only from a pool of rain-water 

 came crawling forth a half -grown, spotted turtle. I 

 must confess to my disappointment. Yet what more 

 suggestive creature could have appeared ? Picking it 

 up, I carried it a dozen paces back into the meadow and 

 placed it in a shallow depression, thickly surrounded by 

 a dense growth of grass. Once upon dry ground, the 

 turtle looked about in every direction, craning its neck 

 to the utmost, aud then turning about, started in great 

 haste directly towards the creek, distant about fifty feet. 

 It could not have taken a straighter course had the stream 

 been in full view. The weeds were high the entire way, 

 and there was no beaten path. How, then, could it so 

 unerringly take the shortest route to the creek ? The 

 waters were not rippling, and so could not be heard, and 

 if the sense of sight availed nothing in the premises, 

 was the creature guided by sense of smell \ This seems 

 probable, but is it not possible that the position of the 

 creek was known to the turtle by the trees that grew 



