CLASSIFICATION OF THE PRIMORDIA 67 



throw new light upon the innumerable details recorded in 

 descriptive literature. 



FIRST EXAMPLE : In the description of Anagallis tenella 

 we find, among other things : segments of the calyx pointed 

 hut short; corolla deeply ^-cleft.^ We recognize here four 

 measurable primordia : (i) length of the calyx ; (2) id. of its 

 segments ; (3) length of the corolla ; (4) id. of its segments. 

 The curve of development of each of these properties may be 

 estabhshed by the study of flower buds. 2 The same primordia 

 may be investigated in the same way through the whole family 

 of the Primulaceae. 



SECOND EXAMPLE : British species of genus Genista ^ : 



Genista : seeds several. 



G. tinctoria : leaves sessile, from narrow lanceolate to broadly 

 elliptical or nearly ovate. Pod nearly an inch long. 



G. pilosa : leaves shortly obovate or lanceolate, obtuse. 

 Pod rather shorter and broader than in G. tinctoria. 



G. anglica : leaves small, lanceolate or ovate. Pods about 

 six lines long, broad. 



The characters mentioned have been observed accurately, 

 but the descriptions are rather vague. Seven measurable 

 properties may be taken into account in each species : (i) 

 length of the petiole * ; (2) id. of the blade ; (3) breadth of the 

 blade ; (4) distance between the greatest breadth of the blade 

 and its summit ; (5) length of the ripe pod ; (6) breadth id. ; 

 (7) number of seeds. ^ 



As long as we confine ourselves to the British flora it seems 

 to be superfluous to examine thoroughly the above-mentioned 

 primordia, because the three British species are easily dis- 

 tinguished by their thorns and hairs. But the quantitative 

 investigation of the leaves, the pods and the seeds would 

 render good service for the description of the numerous species 

 (about seventy) of Genista which are found in other countries 

 and eventually of their subspecies. And, moreover, a descrip- 

 tion ought to be something more than a short diagnosis. 



THIRD EXAMPLE : In certain genera of the family of the 

 CarabidcB the labrum bears two hairs ; in other genera the hairs 

 are four in number. In the description of the species of the 

 genus Carabus these hairs are not mentioned, probably because 

 they are mmierous and their number variable in each species of 

 this genus. I have counted the hairs under consideration in 

 numerous species of Carabus, taking at least six $ and six ? of 



1 HOOKER, British Flora, 1866, p. 306. 



* The turgor being suppressed in one or another way. 

 =» HOOKER, British Flora, i8ry5. pp. 108-109. 



* When the leaves are sessile, the length of the petiole = o. 

 ^ See, on the measurement of variable properties, Part VI. 



