368 THE NATURALIST. 



for what is in tlie nature of tilings. We do not reproacli him with this ; 

 we wonkl only suggest to him the question, that after all these fluctuations, 

 is he 5?M'e that he has seized the real unit, lyerfedhj Junited, constant, and 

 invarlahle, essentially sejxirated every one from the other ? He would 

 answer in the affirmative. For ourselves, we are far from certain of it ; first, 

 because the criterion is defective, and secondly, because we can scarcely 

 believe, sj^ite of all his activity, which is justly admitted, that he has had 

 the time and the prodigious jiatience to malce all the comparisons required 

 for the numerous new types which he proj)oses. We think, that in a great 

 number of cases, he has been content ^^dth analogies, and that he has 

 inferred the general from the particular, a method which frequently leads to 

 error, unless we suppose that all the forms in one genus or one family 

 have the same degree of stability or variability, which is scarcely probable. 

 Take, for example, the genus Erojjhila, (^) which already numbers fifty-three 

 distinct types for France alone, and which, " when all the forms which exist 

 in various parts of this country have been made the subject of attentive 

 stud}^, and compared together in the living state, may reach to more than 

 one-hundred," and imagine what astounding patience must have been required 

 to recognise with certainty the momentary stability of these fifty-three types. 

 Before being enabled to establish them, it was requisite in the begin- 

 ning to make a first division, furnishing two principal groups ; these two 

 had then to be subdivided into two and seven secondary groups. Long 

 and patient comparisons had then to be made, in order to ascertain that 

 the forms in the one group do not pass into the forms of the other, — that 

 those of each secondary group never present the characters of a neighboimng 

 secondary group. Each of the sub-grouj)s must have been separated into 

 species or units. Each unit must have been studied in a sufficient number of 

 examples, and this during several generations. The individuals produced 

 by generation, must have been compared with their parents and their de- 

 scendants. All the organs must have been passed under review in each 

 generation, the compass must have been used to measure the leaves and 

 the internodes, almost to a millimetre, the glass to count the hahs, &c., 

 &c. When we know by what minute characters — almost indiscernible — 

 the units of the genus Eroj)liila are separated from one another, we are amazed 

 at the almost fabulous patience which must have been exercised in making 

 these delicate comparisons upon thousands and thousands of individuals! 

 Besides, in order to arrive at a sure result, these fifty-three types should have 

 (6) " Species 53 sequentes ex Draba verna L. typo." Diagnoses p. 207. [Eds. Nat.] 



