[ 176 ] 



Tn the two laft mixtures the proportions vary confiderably : The 

 firft may ferve as a model for the heavier foils, and the fecond for 

 the lighter. In thefc and the following experiments the carbonic 

 principle feems to have been extraded from the furrounding gar- 

 den mould with which the pots communicated by means of their 

 perforation at bottom. 



Barren Mixtures 



First. 



Mr. Tillet, in his fixth and eighth experiment, mixed \ of 

 potters clay with |- of parings of limeflone and -f- oi Jine fand ; 

 the only difference between this mixture and that of the firft 

 experiment was, that in the firft experiment coarfe fand was 

 ufed and in this Jine, yet the former was fruitful in the higheft 

 degree •, but in this the grain profpered indeed the firft year, but 

 fickened in the fecond, and failed in the third — the proportions 

 have been already ftated. Here we have a clear proof of the ne- 

 celfity of an open texture in foils, without which the beft propor- 

 tions are ufelefs. , 



Second. 



In his thirteenth experiment he employed a mixture of |- potters 

 clay, \ coarfe fand and \ marie. The corn grew well the firft 

 year, poorly the fecond, and decayed the third. The compofition 

 of the marie is not mentioned ; but fuppofing it to contain 70 



per 



