SOURCE OF ST. PETEr's RIVER. 187 



dant to prove injurious to agriculture ; we observed as a 

 distinction between those seen within the two last days, 

 and those met with east of Rock river, that the for- 

 mer contain principally hornblende instead of mica in their 

 composition, while the boulders near Lake Michigan were 

 chiefly granitic. The rock, which has given rise to the 

 Jiornblendic boulders, is one of a peculiar and interest- 

 ing nature ; it differs from sienite by the presence of quartz, 

 from granite by the substitution of hornblende for mica. This 

 rock has not received much attention from European authors ; 

 it does not appear to occupy a very important rank in the ge- 

 ology of Europe, while on the contrary it is very abundant 

 in North America. Those, who are conversant with the 

 mineralogy of New Jersey, know that it constitutes most 

 of the primitive rocks which are found in West Jersey, 

 and which have been described either as granite or sienite ; 

 however extensive that deposite may be, it bears no com- 

 parison to the extensive formation of this rock, which we 

 shall have occasion hereafter to describe, and from which 

 the fragments which constitute the boulders found in Illi- 

 nois, Indiana, Ohio, &c. have, as we believe, been de- 

 tached. 



After travelling eighteen miles, we reached a small stream, 

 designated under the name of Pek-tan-noHs, a diminutive of 

 P^ktannon,* a neighbouring stream into which it discharges 

 itself a few miles below. The meaning of this last in 

 the Sauk language is muddy, and it is remarkable that the 



* As we have had frequent opportunities of observing a nasal termi- 

 nation in Indian words, belonging both to the Sauk, Dacota, and other 

 languages, we have adopted the sign (ii) to designate this sound, which 

 is equivalent to the nasal termination of the French language, thus in 

 the word Pektannon, the last syllable is pronounced by the Indians, 

 exactly as the word non is by the French. 



