ARTIFICIAL CLASSIFICATION, 



95 



Others, as Rivinus (1690), Ludimg' (1750), and Tourne- 

 Jhrt (1700), regard the corolla only in classification ; whilst 

 Rivinus attended to the number and regularity of the parts, 

 Tournefort to the general form of the corolla, and John Ray 

 (1682), connected the fruit with the corolla. All these Co- 

 rollist(E^ as they were named, paid, however, a constant regard 

 to the natural affinities. 



It is scarcely worth while to mention the unsuccessful at- 

 tempts of Antony Magnol (1720), to class plants according 

 to their calyx, and of Sauvage (1751), to arrange them ac- 

 cording to their leaves. We are far from considering every' 

 artificial system as just as good as another, because we find 

 some fault with them all. Our opinion on the contrary is, 

 that the Linnsean system, with the improvements which Smith 

 has proposed, is best adapted to the instruction of beginners. 



CHAP. III. 



ON THE MUTUAL CONNECTIONS OF PLANTS. 



I. Idea of Species. 

 142. 



By Species [species), we understand a number of plants^ 

 which agree with one another in invariable marks. 



In this matter every thing depends upon the idea of inva- 

 riableness. When an organ, or a property of it, is changed 

 neither by diiference of soil, of climate, or of treatment, nor 

 by continued breeding, this organ or property is said to be 

 invariable. When, for instance, we have remarked during 

 centuries, that the Centifolia has always unarmed leaf-stalks, 

 we say correctly, that this property of the Centifolia is inva- 

 riable. 



