Selecting Power of' Plants in regard to Earthy Matters. 165 

 cing them under circumstances so unnatural, and consequently 

 so unfavourable to growth, the same end would be fulfilled if 

 the seeds were sown in some earth, which, though foreign to 

 their constitution, agreed, nevertheless, more nearly in mechani- 

 cal properties with those contained in the soils in which they 

 were wont to grow. It was with this intent that I was origi- 

 nally led to select as a soil for my plants the sulphate of stron- 

 tian, which is obtained in abundance near Bristol, reduced to 

 fine powder: and having found that the ashes of plants which 

 had been reared in this matrix seemed to contain no trace of the 

 earth, I was led, in the next place, to try whether this might be 

 owing merely to the insolubility of the substance in question ; 

 for which reason I varied the experiment by watering my plants 

 with a weak solution of nitrate of strontian. It will appear from 

 the subsequent details, that in either form of the experiment, 



. lime, and not strontites, was the earth that presented itself; but 

 as in proportion to the care that had been taken to exclude any 

 external source of supply for earthy matter, the quantity obtained 

 from the ashes grew less and less, it would be rash to infer, from 

 the small excess of lime which was detected, any power belong- 

 ing to the plant of forming it, when not suppUed from without. 

 Should it, however, appear that a vegetable, which, though not 



, perhaps in full vigour, was at least in a growing and healthy 

 condition, remained in contact with strontian, both in the state 

 of sulphate, and likewise in that of nitrate dissolved in water, 

 for months together, without absorbing any portion, and that, 

 although in want of earthy matter, as its laxity of fibre evidently 

 betrayed, the conclusion would seem to follow, that plants have, 

 to a certain extent, the power, as living agents, of rejecting such 

 substances as, without being poisonous, are unusual to them, and 

 probably unfitted for their economy and structure. Omitting 

 some previous experiments, of which I have preserved no cor- 

 rect notes, I will, in the first instance, refer to one made in 1827, 

 in which grasses and trefoils of various kinds, which had been 

 watered from time to time with a solution of nitrate of strontian, 

 were foimd, on examination, to possess no trace of this earth. In 

 the above instance, however, as the plants had grown in common 

 garden mould, all that could be inferred was, that when lime 

 and strontian are both presented in a state of solution to their 

 roots, they select the former, and reject the latter. In 1829, 



